|
import os |
|
import numpy as np |
|
import gradio as gr |
|
import cv2 |
|
import tensorflow as tf |
|
import keras |
|
from keras.models import Model |
|
from keras.preprocessing import image |
|
|
|
from huggingface_hub import hf_hub_download |
|
import pandas as pd |
|
from PIL import Image |
|
import plotly.express as px |
|
import time |
|
import os |
|
os.environ['TF_ENABLE_ONEDNN_OPTS'] = '0' |
|
|
|
theme = gr.themes.Soft( |
|
primary_hue="purple", |
|
secondary_hue="yellow", |
|
text_size="sm", |
|
) |
|
|
|
css = """ |
|
/* Police lisible et compatible */ |
|
body { |
|
font-family: Arial, sans-serif !important; |
|
} |
|
/* Conteneur du diagnostic global */ |
|
.diagnostic-global { |
|
padding: 10px; |
|
} |
|
/* Style pour les badges Malin/Bénin */ |
|
.highlight.malin { |
|
background-color: #F54927; |
|
color: white; |
|
padding: 4px 8px; |
|
border-radius: 4px; |
|
display: inline-block; |
|
font-weight: bold; |
|
font-size:24px; |
|
margin: 5px 0; |
|
} |
|
.highlight.benin { |
|
background-color: #34EA3A; |
|
color: black; |
|
padding: 4px 8px; |
|
border-radius: 4px; |
|
display: inline-block; |
|
font-weight: bold; |
|
font-size:24px; |
|
margin: 5px 0; |
|
} |
|
/* Conteneur pour le warning */ |
|
.warning-container { |
|
background-color: #f9f9f9; |
|
border-radius: 5px; |
|
padding: 0px; |
|
margin-bottom: 0px; |
|
border: 1px solid #e0e0e0; |
|
} |
|
/* Message d'avertissement */ |
|
.warning-message { |
|
background-color: #e9d5ff; |
|
border-radius: 3px; |
|
padding: 10px; |
|
border: 1px solid #d4b5ff; |
|
font-size: 14px; |
|
color: #333; |
|
font-family: Arial, sans-serif !important; |
|
} |
|
/* Pour les images dans le diagnostic */ |
|
.diagnostic-global img { |
|
max-width: 100%; |
|
height: auto; |
|
float: left; |
|
margin-right: 10px; |
|
margin-bottom: 10px; |
|
} |
|
.feedback-container { |
|
background-color: #f9f9f9; |
|
border-radius: 5px; |
|
padding: 0px; |
|
margin-bottom: 0px; |
|
border: 1px solid #e0e0e0; |
|
font-size:16px; |
|
font-family: Arial, sans-serif !important; |
|
} |
|
/* Clearfix pour les flottants */ |
|
.clearfix::after { |
|
content: ""; |
|
display: table; |
|
clear: both; |
|
} |
|
|
|
""" |
|
|
|
|
|
os.environ['CUDA_VISIBLE_DEVICES'] = '-1' |
|
os.environ['TF_CPP_MIN_LOG_LEVEL'] = '2' |
|
tf.config.set_visible_devices([], 'GPU') |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
from transformers import pipeline |
|
import torch |
|
|
|
|
|
CLASS_NAMES_FULL = { |
|
"mel": "Melanoma", |
|
"bcc": "Basal Cell Carcinoma", |
|
"akiec": "Actinic Keratosis", |
|
"bkl": "Benign Keratosis", |
|
"df": "Dermatofibroma", |
|
"nv": "Melanocytic Nevus", |
|
"vasc": "Vascular Lesion" |
|
} |
|
|
|
|
|
med_nlp = None |
|
try: |
|
print("📥 Chargement du modèle médical BiomedBERT...") |
|
med_nlp = pipeline( |
|
"text-classification", |
|
model="microsoft/BiomedNLP-BiomedBERT-base-uncased-abstract", |
|
framework="pt", |
|
device=0 if torch.cuda.is_available() else -1, |
|
return_all_scores=True |
|
) |
|
print("✅ BiomedBERT chargé avec succès") |
|
except Exception as e: |
|
print(f"❌ Erreur BiomedBERT: {e}") |
|
med_nlp = None |
|
|
|
|
|
|
|
|
|
CLASS_NAMES = ['akiec', 'bcc', 'bkl', 'df', 'nv', 'vasc', 'mel'] |
|
label_to_index = {name: i for i, name in enumerate(CLASS_NAMES)} |
|
diagnosis_map = { |
|
'akiec': 'Bénin', 'bcc': 'Malin', 'bkl': 'Bénin', 'df': 'Bénin', |
|
'nv': 'Bénin', 'vasc': 'Bénin', 'mel': 'Malin' |
|
} |
|
description = { |
|
"akiec": "Bénin : AKIEC ou - kératoses solaires - sont des excroissances précancéreuses provoquées par l'exposition solaire prolongée. Le risque de progression d'une Kératose Actinique vers un carcinome épidermoïde (cancer de la peau non mélanome) existe mais reste modéré", |
|
"bcc": "Malin : BCC ou - carcinome basocellulaire - est un type de cancer cutané. C’est le cancer de la peau le plus fréquent. Il se manifeste par la formation d'une masse, d'un bouton ou d'une lésion sur la couche externe de la peau. Selon Ameli.fr : Le carcinome basocellulaire est la forme la plus courante de cancer de la peau. Ce type de cancer de la peau ne se propage généralement pas, mais nécessite un traitement. Les carcinomes basocellulaires se développent le plus souvent dans les zones d'exposition de la peau au soleil.", |
|
"bkl": "Bénin : BKL ou - kératose séborrhéique - se présente sous la forme d’une lésion pigmentée, en relief, verruqueuse (qui ressemble à une verrue), souvent croûteuse, de plus ou moins grande taille.", |
|
"df": "Bénin : DF ou - dermatofibrome - est une lésion papuleuse ou nodulaire ferme, le plus souvent de petite taille, de couleur rouge marron, de nature fibrohistiocytaire.", |
|
"nv": "Bénin : NV (Nevus) ou les - grains de beauté -, couramment appelés nevus mélanocytaires représentent une accumulation localisée de mélanocytes dans la peau", |
|
"vasc": "Bénin : VASC ou - lésion vasculaire - se traduit par la survenue d’anomalies visibles en surface de la peau et d’aspect variable : rougeurs, taches planes ou en relief, capillaires sanguins apparents", |
|
"mel": "Malin : MEL ou - Mélanome - est un nodule noir ou couleur « peau » présent sur n'importe quelle partie de la peau. Sa consistance est ferme et le nodule peut s'ulcérer, se couvrir d'une croûte, suinter ou saigner. Selon Ameli.fr : Le mélanome est un cancer de la peau peu fréquent, mais grave s'il n'est pas diagnostiqué tôt. L'exposition aux rayons ultraviolets (soleil ou lampe UV) est le principal facteur favorisant sa survenue." |
|
} |
|
|
|
|
|
def load_models_safely(): |
|
models = {} |
|
try: |
|
print("📥 Téléchargement ResNet50...") |
|
resnet_path = hf_hub_download(repo_id="ericjedha/resnet50", filename="Resnet50.keras") |
|
models['resnet50'] = keras.saving.load_model(resnet_path, compile=False) |
|
print("✅ ResNet50 chargé") |
|
except Exception as e: |
|
models['resnet50'] = None |
|
|
|
try: |
|
print("📥 Téléchargement DenseNet201...") |
|
densenet_path = hf_hub_download(repo_id="ericjedha/densenet201", filename="Densenet201.keras") |
|
models['densenet201'] = keras.saving.load_model(densenet_path, compile=False) |
|
print("✅ DenseNet201 chargé") |
|
except Exception as e: |
|
models['densenet201'] = None |
|
|
|
try: |
|
print("📥 Chargement Xception local...") |
|
if os.path.exists("Xception.keras"): |
|
models['xception'] = keras.saving.load_model("Xception.keras", compile=False) |
|
print("✅ Xception chargé") |
|
else: |
|
models['xception'] = None |
|
except Exception as e: |
|
models['xception'] = None |
|
|
|
loaded = {k: v for k, v in models.items() if v is not None} |
|
if not loaded: |
|
raise Exception("❌ Aucun modèle n'a pu être chargé!") |
|
|
|
print(f"🎯 Modèles chargés: {list(loaded.keys())}") |
|
return models |
|
|
|
try: |
|
models_dict = load_models_safely() |
|
model_resnet50 = models_dict.get('resnet50') |
|
model_densenet = models_dict.get('densenet201') |
|
model_xcept = models_dict.get('xception') |
|
except Exception as e: |
|
print(f"🚨 ERREUR CRITIQUE: {e}") |
|
model_resnet50 = model_densenet = model_xcept = None |
|
|
|
|
|
from tensorflow.keras.applications.xception import preprocess_input as preprocess_xception |
|
from tensorflow.keras.applications.resnet50 import preprocess_input as preprocess_resnet |
|
from tensorflow.keras.applications.densenet import preprocess_input as preprocess_densenet |
|
|
|
|
|
def _renorm_safe(p: np.ndarray) -> np.ndarray: |
|
p = np.clip(p, 0.0, None) |
|
s = np.sum(p) |
|
if s <= 0: |
|
return np.ones_like(p, dtype=np.float32) / len(p) |
|
normalized = p / s |
|
return normalized / np.sum(normalized) if np.sum(normalized) > 1.0001 else normalized |
|
|
|
def get_primary_input_name(model): |
|
if isinstance(model.inputs, list) and len(model.inputs) > 0: |
|
return model.inputs[0].name.split(':')[0] |
|
return "input_1" |
|
|
|
def _update_progress(progress, value, desc=""): |
|
""" |
|
Met à jour la barre de progression. |
|
""" |
|
if progress is not None: |
|
progress(value / 100.0, desc=desc) |
|
|
|
|
|
def predict_single(img_input, weights=(0.45, 0.25, 0.3), normalize=True): |
|
print("🔍 DEBUG GRADIO HARMONISÉ - Début de la prédiction") |
|
|
|
|
|
if isinstance(img_input, str): |
|
|
|
img_path = img_input |
|
print(f"📁 Chargement depuis fichier: {img_path}") |
|
|
|
img_raw_x = image.load_img(img_path, target_size=(299, 299)) |
|
img_raw_r = image.load_img(img_path, target_size=(224, 224)) |
|
img_raw_d = image.load_img(img_path, target_size=(224, 224)) |
|
|
|
else: |
|
|
|
print("📁 Chargement depuis upload Gradio") |
|
|
|
|
|
from io import BytesIO |
|
import PIL.Image |
|
|
|
|
|
if img_input.mode != 'RGB': |
|
img_input = img_input.convert('RGB') |
|
print(f"🔄 Conversion en RGB effectuée") |
|
|
|
|
|
img_raw_x = img_input.resize((299, 299), PIL.Image.Resampling.LANCZOS) |
|
img_raw_r = img_input.resize((224, 224), PIL.Image.Resampling.LANCZOS) |
|
img_raw_d = img_input.resize((224, 224), PIL.Image.Resampling.LANCZOS) |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
array_x = image.img_to_array(img_raw_x) |
|
array_r = image.img_to_array(img_raw_r) |
|
array_d = image.img_to_array(img_raw_d) |
|
|
|
print(f"📸 Images loaded:") |
|
print(f" Xception (299x299): {array_x.shape}") |
|
print(f" ResNet (224x224): {array_r.shape}") |
|
print(f" DenseNet (224x224): {array_d.shape}") |
|
|
|
print(f"🔧 Arrays avant preprocessing:") |
|
print(f" X shape: {array_x.shape}, dtype: {array_x.dtype}, range: [{array_x.min()}, {array_x.max()}]") |
|
print(f" R shape: {array_r.shape}, dtype: {array_r.dtype}, range: [{array_r.min()}, {array_r.max()}]") |
|
print(f" D shape: {array_d.shape}, dtype: {array_d.dtype}, range: [{array_d.min()}, {array_d.max()}]") |
|
|
|
|
|
expected_range = (0, 255) |
|
actual_ranges = [ |
|
(array_x.min(), array_x.max()), |
|
(array_r.min(), array_r.max()), |
|
(array_d.min(), array_d.max()) |
|
] |
|
|
|
print(f"🔍 Vérification des ranges:") |
|
for i, (min_val, max_val) in enumerate(actual_ranges): |
|
model_name = ['Xception', 'ResNet', 'DenseNet'][i] |
|
if min_val < expected_range[0] or max_val > expected_range[1]: |
|
print(f" ⚠️ {model_name}: range inhabituel [{min_val}, {max_val}]") |
|
else: |
|
print(f" ✅ {model_name}: range normal [{min_val}, {max_val}]") |
|
|
|
|
|
img_x = np.expand_dims(preprocess_xception(array_x), axis=0) |
|
img_r = np.expand_dims(preprocess_resnet(array_r), axis=0) |
|
img_d = np.expand_dims(preprocess_densenet(array_d), axis=0) |
|
|
|
print(f"🔧 Après preprocessing:") |
|
print(f" Xception range: [{img_x.min():.6f}, {img_x.max():.6f}]") |
|
print(f" ResNet range: [{img_r.min():.6f}, {img_r.max():.6f}]") |
|
print(f" DenseNet range: [{img_d.min():.6f}, {img_d.max():.6f}]") |
|
|
|
|
|
preds = {} |
|
if model_xcept is not None: |
|
preds['xception'] = model_xcept.predict(img_x, verbose=0)[0] |
|
print("\n--- Xception ---") |
|
for i, (class_name, prob) in enumerate(zip(CLASS_NAMES, preds['xception'])): |
|
print(f"{class_name}: {prob*100:.2f}%") |
|
|
|
if model_resnet50 is not None: |
|
preds['resnet50'] = model_resnet50.predict(img_r, verbose=0)[0] |
|
print("\n--- ResNet ---") |
|
for i, (class_name, prob) in enumerate(zip(CLASS_NAMES, preds['resnet50'])): |
|
print(f"{class_name}: {prob*100:.2f}%") |
|
|
|
if model_densenet is not None: |
|
preds['densenet201'] = model_densenet.predict(img_d, verbose=0)[0] |
|
print("\n--- DenseNet ---") |
|
for i, (class_name, prob) in enumerate(zip(CLASS_NAMES, preds['densenet201'])): |
|
print(f"{class_name}: {prob*100:.2f}%") |
|
|
|
|
|
ensemble = np.zeros(len(CLASS_NAMES), dtype=np.float32) |
|
if 'xception' in preds: ensemble += weights[0] * preds['xception'] |
|
if 'resnet50' in preds: ensemble += weights[1] * preds['resnet50'] |
|
if 'densenet201' in preds: ensemble += weights[2] * preds['densenet201'] |
|
|
|
print("\n--- Ensemble avant mel boost ---") |
|
for i, (class_name, prob) in enumerate(zip(CLASS_NAMES, ensemble)): |
|
print(f"{class_name}: {prob*100:.2f}%") |
|
print("Ensemble sum avant mel boost:", np.sum(ensemble)) |
|
|
|
|
|
mel_idx = label_to_index['mel'] |
|
if 'densenet201' in preds: |
|
old_mel_prob = ensemble[mel_idx] |
|
ensemble[mel_idx] = 0.5 * ensemble[mel_idx] + 0.5 * preds['densenet201'][mel_idx] |
|
print(f"\nMel boost: {old_mel_prob*100:.2f}% -> {ensemble[mel_idx]*100:.2f}%") |
|
|
|
print("\n--- Ensemble après mel boost ---") |
|
for i, (class_name, prob) in enumerate(zip(CLASS_NAMES, ensemble)): |
|
print(f"{class_name}: {prob*100:.2f}%") |
|
|
|
if normalize: |
|
ensemble_before_norm = ensemble.copy() |
|
ensemble = _renorm_safe(ensemble) |
|
print("\n--- Ensemble final après normalisation ---") |
|
for i, (class_name, prob) in enumerate(zip(CLASS_NAMES, ensemble)): |
|
print(f"{class_name}: {prob*100:.2f}%") |
|
print("Ensemble sum final:", np.sum(ensemble)) |
|
|
|
preds['ensemble'] = ensemble |
|
return preds |
|
|
|
|
|
|
|
LAST_CONV_LAYERS = { |
|
"xception": "block14_sepconv2_act", |
|
"resnet50": "conv5_block3_out", |
|
"densenet201": "conv5_block32_concat" |
|
} |
|
|
|
def find_last_dense_layer(model): |
|
for layer in reversed(model.layers): |
|
if isinstance(layer, keras.layers.Dense): |
|
return layer |
|
raise ValueError("Aucune couche Dense trouvée dans le modèle.") |
|
|
|
|
|
def make_gradcam(image_pil, model, last_conv_layer_name, class_index, progress=None): |
|
""" |
|
Grad-CAM avec progression fluide grâce aux micro-pauses |
|
""" |
|
if model is None: |
|
return np.array(image_pil) |
|
|
|
try: |
|
steps = [ |
|
(5, "🔄 Initialisation..."), |
|
(10, "🖼️ Analyse de l'image..."), |
|
(15, "⚙️ Configuration du preprocesseur..."), |
|
(20, "📐 Redimensionnement image..."), |
|
(25, "🧠 Configuration du modèle..."), |
|
(30, "🔗 Création du gradient model..."), |
|
(35, "⚡ Préparation du calcul..."), |
|
(40, "🔥 Forward pass..."), |
|
(45, "📊 Calcul des activations..."), |
|
(50, "🎯 Extraction classe cible..."), |
|
(55, "⚡ Calcul du gradient..."), |
|
(60, "📈 Traitement des gradients..."), |
|
(70, "📊 Pooling des gradients..."), |
|
(75, "🎨 Construction heatmap..."), |
|
(80, "🌡️ Normalisation heatmap..."), |
|
(85, "🎯 Application colormap..."), |
|
(90, "🖼️ Redimensionnement final..."), |
|
(95, "✨ Superposition images..."), |
|
(100, "✅ Terminé !") |
|
] |
|
|
|
step = 0 |
|
def next_step(): |
|
nonlocal step |
|
if step < len(steps): |
|
val, desc = steps[step] |
|
_update_progress(progress, val, desc) |
|
time.sleep(0.02) |
|
step += 1 |
|
|
|
next_step() |
|
|
|
|
|
input_size = model.input_shape[1:3] |
|
if 'xception' in model.name.lower(): |
|
preprocessor = preprocess_xception |
|
elif 'resnet50' in model.name.lower(): |
|
preprocessor = preprocess_resnet |
|
elif 'densenet' in model.name.lower(): |
|
preprocessor = preprocess_densenet |
|
else: |
|
preprocessor = preprocess_densenet |
|
|
|
next_step() |
|
next_step() |
|
|
|
|
|
img_np = np.array(image_pil.convert("RGB")) |
|
img_resized = cv2.resize(img_np, input_size) |
|
img_array_preprocessed = preprocessor(np.expand_dims(img_resized, axis=0)) |
|
|
|
next_step() |
|
next_step() |
|
|
|
|
|
try: |
|
conv_layer = model.get_layer(last_conv_layer_name) |
|
except ValueError: |
|
return img_resized |
|
|
|
grad_model = Model(model.inputs, [conv_layer.output, model.output]) |
|
input_name = get_primary_input_name(model) |
|
input_for_model = {input_name: img_array_preprocessed} |
|
|
|
next_step() |
|
next_step() |
|
next_step() |
|
|
|
|
|
with tf.GradientTape() as tape: |
|
next_step() |
|
|
|
last_conv_layer_output, preds = grad_model(input_for_model, training=False) |
|
|
|
next_step() |
|
|
|
if isinstance(preds, list): |
|
preds = preds[0] |
|
class_channel = preds[:, int(class_index)] |
|
|
|
next_step() |
|
|
|
grads = tape.gradient(class_channel, last_conv_layer_output) |
|
if grads is None: |
|
return img_resized |
|
|
|
next_step() |
|
next_step() |
|
|
|
|
|
pooled_grads = tf.reduce_mean(grads, axis=(0, 1, 2)) |
|
last_conv_layer_output = last_conv_layer_output[0] |
|
|
|
next_step() |
|
|
|
|
|
heatmap = last_conv_layer_output @ pooled_grads[..., tf.newaxis] |
|
heatmap = tf.squeeze(heatmap) |
|
heatmap = tf.maximum(heatmap, 0) |
|
max_val = tf.math.reduce_max(heatmap) |
|
|
|
if max_val == 0: |
|
heatmap = tf.ones_like(heatmap) * 0.5 |
|
else: |
|
heatmap = heatmap / max_val |
|
|
|
next_step() |
|
next_step() |
|
|
|
|
|
heatmap_np = heatmap.numpy() |
|
heatmap_np = np.clip(heatmap_np.astype(np.float32), 0, 1) |
|
|
|
heatmap_resized = cv2.resize(heatmap_np, (img_resized.shape[1], img_resized.shape[0])) |
|
heatmap_uint8 = np.uint8(255 * heatmap_resized) |
|
heatmap_colored = cv2.applyColorMap(heatmap_uint8, cv2.COLORMAP_JET) |
|
|
|
next_step() |
|
next_step() |
|
|
|
|
|
img_bgr = cv2.cvtColor(img_resized, cv2.COLOR_RGB2BGR) |
|
superimposed_img = cv2.addWeighted(img_bgr, 0.6, heatmap_colored, 0.4, 0) |
|
|
|
next_step() |
|
|
|
return cv2.cvtColor(superimposed_img, cv2.COLOR_BGR2RGB) |
|
|
|
except Exception as e: |
|
import traceback |
|
traceback.print_exc() |
|
_update_progress(progress, 100, "❌ Erreur") |
|
return np.array(image_pil) |
|
|
|
|
|
|
|
|
|
|
|
def get_medical_analysis_simple(top_class_name, confidence, mel_prob): |
|
""" |
|
Analyse médicale simplifiée et claire |
|
""" |
|
global med_nlp |
|
|
|
if med_nlp is None: |
|
return get_fallback_analysis(top_class_name, confidence, mel_prob) |
|
|
|
try: |
|
|
|
medical_text = f""" |
|
Skin lesion diagnosis: {CLASS_NAMES_FULL.get(top_class_name, top_class_name)} |
|
Diagnostic confidence: {confidence:.0f}% |
|
Melanoma risk: {mel_prob:.1f}% |
|
Clinical assessment required for {top_class_name} lesion evaluation. |
|
""" |
|
|
|
|
|
result = med_nlp(medical_text) |
|
print("DEBUG RESULT:", result) |
|
|
|
if result and len(result) > 0: |
|
|
|
candidates = result[0] if isinstance(result[0], list) else result |
|
|
|
|
|
best_score = max(candidates, key=lambda x: x['score'])['score'] |
|
|
|
|
|
bert_assessment = get_simple_bert_assessment(best_score, top_class_name, confidence, mel_prob) |
|
|
|
return { |
|
"status": "success", |
|
"assessment": bert_assessment['message'], |
|
"urgency": bert_assessment['urgency'], |
|
"recommendation": bert_assessment['recommendation'] |
|
} |
|
else: |
|
return get_fallback_analysis(top_class_name, confidence, mel_prob) |
|
|
|
except Exception as e: |
|
print(f"Erreur BiomedBERT: {e}") |
|
return get_fallback_analysis(top_class_name, confidence, mel_prob) |
|
|
|
|
|
def get_simple_bert_assessment(bert_score, class_name, confidence, mel_prob): |
|
""" |
|
Assessment cohérent SkinAI + BiomedBERT |
|
""" |
|
|
|
if confidence >= 95: |
|
message = "✅ Diagnostic confirmé - forte confiance du modèle principal" |
|
urgency_modifier = "" |
|
|
|
|
|
elif confidence >= 70: |
|
if bert_score > 0.6: |
|
message = "👍 Diagnostic probable - SkinAI et BiomedBERT convergent" |
|
urgency_modifier = "" |
|
else: |
|
message = "⚠️ Diagnostic possible - SkinAI confiant mais BiomedBERT reste incertain" |
|
urgency_modifier = " (évaluation complémentaire recommandée)" |
|
|
|
|
|
else: |
|
if bert_score > 0.6: |
|
message = "🤝 Diagnostic appuyé par BiomedBERT malgré faible confiance SkinAI" |
|
urgency_modifier = "" |
|
else: |
|
message = "🔍 Diagnostic incertain - ni SkinAI ni BiomedBERT ne sont sûrs" |
|
urgency_modifier = " (consultation urgente pour clarification)" |
|
|
|
|
|
base_urgency = get_base_urgency_simple(class_name, mel_prob) |
|
|
|
|
|
if confidence < 70 and bert_score <= 0.6: |
|
if "routine" in base_urgency.lower(): |
|
urgency = "Consultation sous 2 semaines" |
|
elif "semaines" in base_urgency: |
|
urgency = "Consultation sous 1 semaine" |
|
else: |
|
urgency = base_urgency |
|
else: |
|
urgency = base_urgency |
|
|
|
|
|
recommendation = get_simple_recommendation(class_name, mel_prob, bert_score < 0.6 and confidence < 95) |
|
|
|
return { |
|
"message": message, |
|
"urgency": urgency + urgency_modifier, |
|
"recommendation": recommendation |
|
} |
|
|
|
|
|
def get_base_urgency_simple(class_name, mel_prob): |
|
""" |
|
Urgence simple selon pathologie |
|
""" |
|
if class_name == "mel" or mel_prob > 20: |
|
return "Consultation immédiate (24-48h)" |
|
elif class_name in ["bcc", "akiec"] or mel_prob > 10: |
|
return "Consultation sous 2-3 semaines" |
|
elif mel_prob > 5: |
|
return "Consultation sous 1 mois" |
|
else: |
|
return "Surveillance de routine" |
|
|
|
def get_simple_recommendation(class_name, mel_prob, bert_uncertain): |
|
""" |
|
Recommandation simple et claire |
|
""" |
|
recommendations = { |
|
"mel": "🚨 **URGENT** : Rendez-vous dermatologue immédiat pour suspicion de mélanome", |
|
"bcc": "⚠️ **Important** : Consultation dermatologue pour traitement du carcinome", |
|
"akiec": "📋 **Surveillance** : Suivi dermatologique pour lésion précancéreuse", |
|
"bkl": "✅ **Bénin** : Surveillance habituelle, pas d'urgence", |
|
"df": "📋 **Suivi** : Consultation si changement d'aspect", |
|
"nv": "✅ **Routine** : Surveillance standard des grains de beauté", |
|
"vasc": "📋 **Évaluation** : Consultation pour caractérisation complète" |
|
} |
|
|
|
base_rec = recommendations.get(class_name, "📋 Consultation dermatologique recommandée") |
|
|
|
|
|
if mel_prob > 15: |
|
base_rec += f"\n⚠️ **Attention** : Risque mélanome élevé ({mel_prob:.0f}%)" |
|
elif mel_prob > 8: |
|
base_rec += f"\n📊 Risque mélanome modéré ({mel_prob:.0f}%)" |
|
|
|
|
|
if bert_uncertain: |
|
base_rec += "\n🔍 **Note** : Présentation atypique détectée, avis spécialisé recommandé" |
|
|
|
return base_rec |
|
|
|
def get_fallback_analysis(class_name, confidence, mel_prob): |
|
""" |
|
Analyse de secours si BiomedBERT indisponible |
|
""" |
|
return { |
|
"status": "fallback", |
|
"assessment": "📋 Analyse standard (BiomedBERT indisponible)", |
|
"urgency": get_base_urgency_simple(class_name, mel_prob), |
|
"recommendation": get_simple_recommendation(class_name, mel_prob, False) |
|
} |
|
|
|
|
|
|
|
|
|
|
|
|
|
current_image = None |
|
current_predictions = None |
|
|
|
|
|
import plotly.graph_objects as go |
|
import numpy as np |
|
import gradio as gr |
|
|
|
def quick_predict_ui(image_pil): |
|
global current_image, current_predictions |
|
if image_pil is None: |
|
empty_fig = go.Figure() |
|
empty_fig.update_layout( |
|
title="Veuillez uploader une image pour voir les probabilités", |
|
height=450, |
|
template="plotly_white", |
|
showlegend=False |
|
) |
|
|
|
return ( |
|
'<div class="diagnostic-global"><h2>Veuillez uploader une image.</h2></div>', |
|
"", |
|
gr.update(value="", visible=False), |
|
empty_fig, |
|
"❌ Erreur: Aucune image fournie.", |
|
"" |
|
) |
|
|
|
try: |
|
current_image = image_pil |
|
all_preds = predict_single(image_pil) |
|
current_predictions = all_preds |
|
ensemble_probs = all_preds["ensemble"] |
|
top_class_idx = int(np.argmax(ensemble_probs)) |
|
top_class_name = CLASS_NAMES[top_class_idx] |
|
global_diag = diagnosis_map[top_class_name] |
|
|
|
mel_idx = CLASS_NAMES.index("mel") |
|
mel_prob = ensemble_probs[mel_idx] * 100 |
|
confidence = ensemble_probs[top_class_idx] * 100 |
|
|
|
|
|
medical_analysis = get_medical_analysis_simple(top_class_name, confidence, mel_prob) |
|
|
|
desc_top = description.get(top_class_name, "") |
|
|
|
|
|
|
|
if global_diag == "Malin": |
|
global_diag_html = f''' |
|
<div class="diagnostic-global clearfix"> |
|
<img src="https://huggingface.co/spaces/ericjedha/skin_care/resolve/main/mel.webp" width="150" style="float:left;margin-right:10px;"> |
|
<div> |
|
<span style="font-size:16px;font-weight:bold;">Diagnostic Global</span><br> |
|
<div class="highlight malin"> |
|
{global_diag} : {ensemble_probs[top_class_idx]*100:.2f}% ▪ {top_class_name.upper()}* |
|
</div> |
|
</div> |
|
</div> |
|
''' |
|
else: |
|
img_src = "non-mel.webp" if global_diag == "Bénin" else "mel.webp" |
|
global_diag_html = f''' |
|
<div class="diagnostic-global clearfix"> |
|
<img src="https://huggingface.co/spaces/ericjedha/skin_care/resolve/main/{img_src}" width="150" style="float:left;margin-right:10px;"> |
|
<div> |
|
<span style="font-size:16px;font-weight:bold;">Diagnostic Global</span><br> |
|
<div class="highlight benin"> |
|
{global_diag} : {ensemble_probs[top_class_idx]*100:.2f}% ▪ {top_class_name.upper()}* |
|
</div> |
|
</div> |
|
</div> |
|
''' |
|
|
|
output_text_html = f'<div class="warning-message"><strong>* Explication du résultat</strong> : {desc_top}</div>' |
|
|
|
warning_visible = False |
|
warning_html = "" |
|
if mel_prob > 5 and top_class_name != "mel": |
|
warning_html = f''' |
|
<div class="warning-message"> |
|
<img src="https://huggingface.co/spaces/ericjedha/skin_care/resolve/main/mel-modere.webp" width="150" style = "float:left"> Le modèle a détecté un risque modéré de mélanome <strong>({mel_prob:.1f}%)</strong>. |
|
<strong>Veuillez consulter votre médecin pour lever tout doute</strong>. |
|
</div> |
|
''' |
|
warning_visible = True |
|
|
|
|
|
biomed_html = format_simple_medical_html(medical_analysis) |
|
|
|
|
|
probabilities = [round(ensemble_probs[i] * 100, 2) for i in range(len(CLASS_NAMES))] |
|
colors = ['#ff6b6b' if name == top_class_name else '#4ecdc4' for name in CLASS_NAMES] |
|
|
|
fig = go.Figure(data=[ |
|
go.Bar( |
|
x=CLASS_NAMES, |
|
y=probabilities, |
|
text=[f'{p:.2f}%' for p in probabilities], |
|
textposition='outside', |
|
marker_color=colors, |
|
hovertemplate='<b>%{x}</b><br>Probabilité: %{y:.2f}%<extra></extra>' |
|
) |
|
]) |
|
|
|
fig.update_layout( |
|
xaxis_title="Classes", |
|
yaxis_title="Probabilité (%)", |
|
yaxis=dict(range=[0, max(probabilities) * 1.15]), |
|
height=450, |
|
template="plotly_white", |
|
showlegend=False, |
|
font=dict(size=12), |
|
margin=dict(l=50, r=50, t=70, b=100) |
|
) |
|
|
|
fig.update_xaxes(tickangle=45) |
|
|
|
return ( |
|
global_diag_html, |
|
output_text_html, |
|
gr.update(value=warning_html, visible=warning_visible), |
|
fig, |
|
"✅ Analyse terminée.", |
|
biomed_html |
|
) |
|
|
|
except Exception as e: |
|
error_fig = go.Figure() |
|
error_fig.update_layout( |
|
title=f"Erreur lors de la création du graphique: {str(e)}", |
|
height=450, |
|
template="plotly_white", |
|
showlegend=False |
|
) |
|
|
|
error_biomed_html = f''' |
|
<div class="medical-analysis" style="background: #ffebee; color: #d32f2f; padding: 15px; border-radius: 10px; border: 1px solid #ef9a9a;"> |
|
<h3>🧬 Analyse Médicale</h3> |
|
<p>❌ Erreur lors de l'analyse</p> |
|
</div> |
|
''' |
|
|
|
return ( |
|
f'<div class="diagnostic-global"><h2>Erreur: {str(e)}</h2></div>', |
|
"", |
|
gr.update(value=f''' |
|
<div class="warning-message" style="background-color:#ffebee;border:1px solid #ef9a9a;"> |
|
❌ Une erreur est survenue : {str(e)} |
|
</div> |
|
''', visible=True), |
|
error_fig, |
|
f"❌ Erreur: {str(e)}", |
|
error_biomed_html |
|
) |
|
|
|
def format_simple_medical_html(medical_analysis): |
|
""" |
|
HTML simple et clair pour l'analyse médicale |
|
""" |
|
|
|
if medical_analysis['status'] == 'error': |
|
return f''' |
|
<div class="medical-analysis" style="background: #ffebee; color: #d32f2f; padding: 15px; border-radius: 10px;"> |
|
<h3>🧬 Analyse Médicale</h3> |
|
<p>❌ Erreur lors de l'analyse</p> |
|
</div> |
|
''' |
|
|
|
|
|
if "✅" in medical_analysis['assessment']: |
|
bg_color = "linear-gradient(135deg, #4caf50 0%, #66bb6a 100%)" |
|
elif "⚠️" in medical_analysis['assessment']: |
|
bg_color = "linear-gradient(135deg, #ff9800 0%, #ffb74d 100%)" |
|
elif "🔍" in medical_analysis['assessment']: |
|
bg_color = "linear-gradient(135deg, #f44336 0%, #ef5350 100%)" |
|
else: |
|
bg_color = "linear-gradient(135deg, #1976d2 0%, #42a5f5 100%)" |
|
|
|
html = f''' |
|
<div class="medical-analysis" style="background: {bg_color}; color: white; padding: 20px; border-radius: 12px; margin-top: 15px; box-shadow: 0 4px 8px rgba(0,0,0,0.2);"> |
|
<h3 style="margin: 0 0 15px 0; font-size: 18px;"> |
|
🤖 BiomedBERT : 2<sup>e<sup> avis |
|
</h3> |
|
|
|
<div style="background: rgba(255,255,255,0.15); padding: 15px; border-radius: 8px; margin-bottom: 15px;"> |
|
<strong>🎯 Évaluation IA :</strong><br> |
|
<div style="margin-top: 8px; font-size: 16px;">{medical_analysis['assessment']}</div> |
|
</div> |
|
|
|
<div style="background: rgba(255,255,255,0.15); padding: 15px; border-radius: 8px; margin-bottom: 15px;"> |
|
<strong>⏰ Délai recommandé :</strong><br> |
|
<div style="margin-top: 8px; font-size: 16px; font-weight: bold;">{medical_analysis['urgency']}</div> |
|
</div> |
|
|
|
<div style="background: rgba(255,255,255,0.2); padding: 15px; border-radius: 8px; border-left: 4px solid #ffd700;"> |
|
<strong>💡 Recommandations :</strong><br> |
|
<div style="margin-top: 10px; line-height: 1.6; white-space: pre-line;">{medical_analysis['recommendation']}</div> |
|
</div> |
|
</div> |
|
''' |
|
|
|
return html |
|
|
|
def generate_gradcam_ui(progress=gr.Progress()): |
|
global current_image, current_predictions |
|
if current_image is None or current_predictions is None: |
|
return None, "❌ Aucun résultat précédent — lance d'abord l'analyse rapide." |
|
|
|
try: |
|
ensemble_probs = current_predictions["ensemble"] |
|
top_class_idx = int(np.argmax(ensemble_probs)) |
|
|
|
|
|
candidates = [] |
|
if model_xcept is not None: |
|
candidates.append(("xception", model_xcept, current_predictions["xception"][top_class_idx])) |
|
if model_resnet50 is not None: |
|
candidates.append(("resnet50", model_resnet50, current_predictions["resnet50"][top_class_idx])) |
|
if model_densenet is not None: |
|
candidates.append(("densenet201", model_densenet, current_predictions["densenet201"][top_class_idx])) |
|
|
|
if not candidates: |
|
return None, "❌ Aucun modèle disponible pour Grad-CAM." |
|
|
|
|
|
explainer_model_name, explainer_model, conf = max(candidates, key=lambda t: t[2]) |
|
explainer_layer = LAST_CONV_LAYERS.get(explainer_model_name) |
|
|
|
|
|
gradcam_img = make_gradcam( |
|
current_image, |
|
explainer_model, |
|
explainer_layer, |
|
class_index=top_class_idx, |
|
progress=progress |
|
) |
|
|
|
return gradcam_img, f"✅ Grad-CAM généré avec {explainer_model_name} (confiance: {conf:.1%})" |
|
|
|
except Exception as e: |
|
import traceback |
|
traceback.print_exc() |
|
return None, f"❌ Erreur: {e}" |
|
|
|
|
|
|
|
|
|
example_paths = ["ISIC_0024627.jpg", "ISIC_0025539.jpg", "ISIC_0031410.jpg"] |
|
|
|
import pandas as pd |
|
import gradio as gr |
|
|
|
with gr.Blocks(theme=theme, title="Analyse de lésions", css=css) as demo: |
|
gr.Markdown("# 🔬 Skin Care : analyse de lésions cutanées") |
|
|
|
models_status = [] |
|
if model_resnet50: |
|
models_status.append("☑ ResNet50") |
|
if model_densenet: |
|
models_status.append("☑ DenseNet201") |
|
if model_xcept: |
|
models_status.append("☑ Xception") |
|
|
|
gr.Markdown(f"**Avertissement 🚨** cette application est un projet d'étudiant. **Seul votre médecin est habilité à vous donner un diagnostic**.") |
|
gr.Markdown(f"**Étape 1️⃣** L'analyse rapide vous donnera le diagnostic global en 10s. **Étape 2️⃣**: le rendu de la carte colorée de votre lésion peut pendre jusqu'à 60s.") |
|
|
|
|
|
with gr.Row(): |
|
with gr.Column(scale=1): |
|
input_image = gr.Image(type="pil", label="📸 Uploader une image") |
|
|
|
with gr.Row(): |
|
quick_btn = gr.Button("Analyse Rapide", variant="primary") |
|
gradcam_btn = gr.Button("Carte colorée (option)", variant="secondary") |
|
|
|
gr.Examples(examples=example_paths, inputs=input_image) |
|
output_gradcam = gr.Image(label="🔍 Carte Colorée Grad-CAM") |
|
output_status = gr.Textbox(label="Statut", interactive=False) |
|
|
|
|
|
with gr.Column(scale=2): |
|
output_label = gr.HTML( |
|
value='<h3 class="output-class">Pour obtenir un diagnostic, uploadez une image ou prenez une photo.</h3>', |
|
elem_classes="diagnostic-global" |
|
) |
|
|
|
|
|
output_text = gr.HTML( |
|
value="", |
|
elem_classes="warning-container" |
|
) |
|
|
|
output_warning = gr.HTML( |
|
value="", |
|
elem_classes="warning-container", |
|
visible=False |
|
) |
|
|
|
|
|
|
|
initial_df = pd.DataFrame({ |
|
'Classes': CLASS_NAMES, |
|
'Probabilités (%)': [0] * len(CLASS_NAMES) |
|
}) |
|
|
|
|
|
output_plot = gr.Plot(label="Probabilités par classe") |
|
|
|
output_medical = gr.HTML(label="🧬 Analyse Médicale Avancée") |
|
|
|
gr.HTML(value="""<div><hr style="width:50%; border-top: 2px #a855f7; height: 10px; border-style:dotted; margin: auto;"></div>""") |
|
gr.Markdown(f"Ensemble de modèles utilisés : {', '.join(models_status) if models_status else 'AUCUN'}") |
|
gr.HTML(value=""" |
|
|
|
<strong>☷ Dataset utilisé</strong> pour l'entrainement des modèles SkinAI : HAM10000, ce dataset HAM10000 a été créé par une équipe internationale dirigée par des chercheurs autrichiens, allemands et australiens. <br> |
|
<strong> 🤖 Les modèles de Machine Learning SkinAI : </strong> <em> 🤖 Xception</em> - Réseau de convolution profond (CNN), <em> 🤖 ResNet50</em> - Réseau résiduel (Residual Network), 🤖 DenseNet201 - Réseau dense (Dense Convolutional Network). <br> |
|
<strong> 🤖 BiomedBERT : </strong> : Microsoft BiomedBERT est un modèle de langage spécialisé dans le domaine biomédical, développé par Microsoft. Il vient en soutient pour des conseils sur les données de la prédiction. <br> |
|
<strong>🏛️ RGPD & Digital Act </strong> : |
|
Ce dataset ne peut pas être utilisé pour des cas réels aujourd'hui notamment du fait qu'il ne comporte qu'essentiellement des peaux de populations européennes (allemands et autrichiens). <br>Cette application ne collecte pas vos données personnelles. <b>Les images uploadées ne sont pas stockées</b>. <br>La politique de Cookies 🍪 est gérée par <a href='https://huggingface.co/privacy'>Hugging Face disponible ici</a>. |
|
""") |
|
|
|
|
|
quick_btn.click( |
|
fn=quick_predict_ui, |
|
inputs=input_image, |
|
outputs=[output_label, output_text, output_warning, output_plot, output_status, output_medical], |
|
) |
|
|
|
gradcam_btn.click( |
|
fn=generate_gradcam_ui, |
|
inputs=[], |
|
outputs=[output_gradcam, output_status] |
|
) |
|
|
|
demo.launch() |