Update README.md
Browse files
README.md
CHANGED
@@ -1,3 +1,344 @@
|
|
1 |
-
---
|
2 |
-
license: mit
|
3 |
-
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: mit
|
3 |
+
---
|
4 |
+
# Evaluation Report for rtdetrx_bb_detect_model
|
5 |
+
|
6 |
+
**Tasks:** Single-class Object Detection, Feature extraction
|
7 |
+
|
8 |
+
## Evaluation Notes
|
9 |
+
|
10 |
+
This model does not identify individual species but detects a single category of object.
|
11 |
+
|
12 |
+
The evaluation was performed on a single-class basis using the text prompt: **'bark_beetle'**.
|
13 |
+
|
14 |
+
### Mantel Correlation Explanation
|
15 |
+
|
16 |
+
The Mantel R statistic is calculated by comparing the distances between clustering centroids of different species to their phylogenetic distances. This helps determine if the model's learned feature representations correlate with the evolutionary relationships between species.
|
17 |
+
|
18 |
+
## Object Classification Performance
|
19 |
+
|
20 |
+
**mAP@[.5:.95]:** 0.921
|
21 |
+
|
22 |
+
### mAP per IoU Threshold
|
23 |
+
|
24 |
+
| IoU Threshold | mAP |
|
25 |
+
|:----------------|---------:|
|
26 |
+
| [email protected] | 0.987815 |
|
27 |
+
| [email protected] | 0.985755 |
|
28 |
+
| [email protected] | 0.983646 |
|
29 |
+
| [email protected] | 0.979705 |
|
30 |
+
| [email protected] | 0.972081 |
|
31 |
+
| [email protected] | 0.956602 |
|
32 |
+
| [email protected] | 0.919432 |
|
33 |
+
| [email protected] | 0.873082 |
|
34 |
+
| [email protected] | 0.824972 |
|
35 |
+
| [email protected] | 0.722574 |
|
36 |
+
|
37 |
+
### Average Precision per Class (at last IoU threshold)
|
38 |
+
|
39 |
+
| Class | AP |
|
40 |
+
|:------------|---------:|
|
41 |
+
| bark_beetle | 0.722574 |
|
42 |
+
|
43 |
+
### Classification Metrics per IoU Threshold
|
44 |
+
|
45 |
+
#### IoU Threshold: iou_0.50
|
46 |
+
|
47 |
+
- **Accuracy:** 0.995
|
48 |
+
- **Balanced Accuracy:** 0.995
|
49 |
+
- **Macro Precision:** 1.000
|
50 |
+
- **Macro Recall:** 0.995
|
51 |
+
- **Macro F1 Score:** 0.997
|
52 |
+
- **Cohen's Kappa:** 0.000
|
53 |
+
- **Matthews Corrcoef:** 0.000
|
54 |
+
|
55 |
+
##### Confusion Matrix
|
56 |
+
|
57 |
+
```
|
58 |
+
Predicted Label 0
|
59 |
+
True Label
|
60 |
+
0 16395
|
61 |
+
```
|
62 |
+
|
63 |
+
##### Classification Report
|
64 |
+
|
65 |
+
```
|
66 |
+
precision recall f1-score support
|
67 |
+
0 1.0 0.994842 0.997414 16480.0
|
68 |
+
micro avg 1.0 0.994842 0.997414 16480.0
|
69 |
+
macro avg 1.0 0.994842 0.997414 16480.0
|
70 |
+
weighted avg 1.0 0.994842 0.997414 16480.0
|
71 |
+
```
|
72 |
+
|
73 |
+
#### IoU Threshold: iou_0.55
|
74 |
+
|
75 |
+
- **Accuracy:** 0.993
|
76 |
+
- **Balanced Accuracy:** 0.993
|
77 |
+
- **Macro Precision:** 1.000
|
78 |
+
- **Macro Recall:** 0.993
|
79 |
+
- **Macro F1 Score:** 0.996
|
80 |
+
- **Cohen's Kappa:** 0.000
|
81 |
+
- **Matthews Corrcoef:** 0.000
|
82 |
+
|
83 |
+
##### Confusion Matrix
|
84 |
+
|
85 |
+
```
|
86 |
+
Predicted Label 0
|
87 |
+
True Label
|
88 |
+
0 16360
|
89 |
+
```
|
90 |
+
|
91 |
+
##### Classification Report
|
92 |
+
|
93 |
+
```
|
94 |
+
precision recall f1-score support
|
95 |
+
0 1.0 0.992718 0.996346 16480.0
|
96 |
+
micro avg 1.0 0.992718 0.996346 16480.0
|
97 |
+
macro avg 1.0 0.992718 0.996346 16480.0
|
98 |
+
weighted avg 1.0 0.992718 0.996346 16480.0
|
99 |
+
```
|
100 |
+
|
101 |
+
#### IoU Threshold: iou_0.60
|
102 |
+
|
103 |
+
- **Accuracy:** 0.990
|
104 |
+
- **Balanced Accuracy:** 0.990
|
105 |
+
- **Macro Precision:** 1.000
|
106 |
+
- **Macro Recall:** 0.990
|
107 |
+
- **Macro F1 Score:** 0.995
|
108 |
+
- **Cohen's Kappa:** 0.000
|
109 |
+
- **Matthews Corrcoef:** 0.000
|
110 |
+
|
111 |
+
##### Confusion Matrix
|
112 |
+
|
113 |
+
```
|
114 |
+
Predicted Label 0
|
115 |
+
True Label
|
116 |
+
0 16323
|
117 |
+
```
|
118 |
+
|
119 |
+
##### Classification Report
|
120 |
+
|
121 |
+
```
|
122 |
+
precision recall f1-score support
|
123 |
+
0 1.0 0.990473 0.995214 16480.0
|
124 |
+
micro avg 1.0 0.990473 0.995214 16480.0
|
125 |
+
macro avg 1.0 0.990473 0.995214 16480.0
|
126 |
+
weighted avg 1.0 0.990473 0.995214 16480.0
|
127 |
+
```
|
128 |
+
|
129 |
+
#### IoU Threshold: iou_0.65
|
130 |
+
|
131 |
+
- **Accuracy:** 0.986
|
132 |
+
- **Balanced Accuracy:** 0.986
|
133 |
+
- **Macro Precision:** 1.000
|
134 |
+
- **Macro Recall:** 0.986
|
135 |
+
- **Macro F1 Score:** 0.993
|
136 |
+
- **Cohen's Kappa:** 0.000
|
137 |
+
- **Matthews Corrcoef:** 0.000
|
138 |
+
|
139 |
+
##### Confusion Matrix
|
140 |
+
|
141 |
+
```
|
142 |
+
Predicted Label 0
|
143 |
+
True Label
|
144 |
+
0 16255
|
145 |
+
```
|
146 |
+
|
147 |
+
##### Classification Report
|
148 |
+
|
149 |
+
```
|
150 |
+
precision recall f1-score support
|
151 |
+
0 1.0 0.986347 0.993127 16480.0
|
152 |
+
micro avg 1.0 0.986347 0.993127 16480.0
|
153 |
+
macro avg 1.0 0.986347 0.993127 16480.0
|
154 |
+
weighted avg 1.0 0.986347 0.993127 16480.0
|
155 |
+
```
|
156 |
+
|
157 |
+
#### IoU Threshold: iou_0.70
|
158 |
+
|
159 |
+
- **Accuracy:** 0.979
|
160 |
+
- **Balanced Accuracy:** 0.979
|
161 |
+
- **Macro Precision:** 1.000
|
162 |
+
- **Macro Recall:** 0.979
|
163 |
+
- **Macro F1 Score:** 0.989
|
164 |
+
- **Cohen's Kappa:** 0.000
|
165 |
+
- **Matthews Corrcoef:** 0.000
|
166 |
+
|
167 |
+
##### Confusion Matrix
|
168 |
+
|
169 |
+
```
|
170 |
+
Predicted Label 0
|
171 |
+
True Label
|
172 |
+
0 16136
|
173 |
+
```
|
174 |
+
|
175 |
+
##### Classification Report
|
176 |
+
|
177 |
+
```
|
178 |
+
precision recall f1-score support
|
179 |
+
0 1.0 0.979126 0.989453 16480.0
|
180 |
+
micro avg 1.0 0.979126 0.989453 16480.0
|
181 |
+
macro avg 1.0 0.979126 0.989453 16480.0
|
182 |
+
weighted avg 1.0 0.979126 0.989453 16480.0
|
183 |
+
```
|
184 |
+
|
185 |
+
#### IoU Threshold: iou_0.75
|
186 |
+
|
187 |
+
- **Accuracy:** 0.965
|
188 |
+
- **Balanced Accuracy:** 0.965
|
189 |
+
- **Macro Precision:** 1.000
|
190 |
+
- **Macro Recall:** 0.965
|
191 |
+
- **Macro F1 Score:** 0.982
|
192 |
+
- **Cohen's Kappa:** 0.000
|
193 |
+
- **Matthews Corrcoef:** 0.000
|
194 |
+
|
195 |
+
##### Confusion Matrix
|
196 |
+
|
197 |
+
```
|
198 |
+
Predicted Label 0
|
199 |
+
True Label
|
200 |
+
0 15902
|
201 |
+
```
|
202 |
+
|
203 |
+
##### Classification Report
|
204 |
+
|
205 |
+
```
|
206 |
+
precision recall f1-score support
|
207 |
+
0 1.0 0.964927 0.982151 16480.0
|
208 |
+
micro avg 1.0 0.964927 0.982151 16480.0
|
209 |
+
macro avg 1.0 0.964927 0.982151 16480.0
|
210 |
+
weighted avg 1.0 0.964927 0.982151 16480.0
|
211 |
+
```
|
212 |
+
|
213 |
+
#### IoU Threshold: iou_0.80
|
214 |
+
|
215 |
+
- **Accuracy:** 0.931
|
216 |
+
- **Balanced Accuracy:** 0.931
|
217 |
+
- **Macro Precision:** 1.000
|
218 |
+
- **Macro Recall:** 0.931
|
219 |
+
- **Macro F1 Score:** 0.964
|
220 |
+
- **Cohen's Kappa:** 0.000
|
221 |
+
- **Matthews Corrcoef:** 0.000
|
222 |
+
|
223 |
+
##### Confusion Matrix
|
224 |
+
|
225 |
+
```
|
226 |
+
Predicted Label 0
|
227 |
+
True Label
|
228 |
+
0 15338
|
229 |
+
```
|
230 |
+
|
231 |
+
##### Classification Report
|
232 |
+
|
233 |
+
```
|
234 |
+
precision recall f1-score support
|
235 |
+
0 1.0 0.930704 0.964108 16480.0
|
236 |
+
micro avg 1.0 0.930704 0.964108 16480.0
|
237 |
+
macro avg 1.0 0.930704 0.964108 16480.0
|
238 |
+
weighted avg 1.0 0.930704 0.964108 16480.0
|
239 |
+
```
|
240 |
+
|
241 |
+
#### IoU Threshold: iou_0.85
|
242 |
+
|
243 |
+
- **Accuracy:** 0.888
|
244 |
+
- **Balanced Accuracy:** 0.888
|
245 |
+
- **Macro Precision:** 1.000
|
246 |
+
- **Macro Recall:** 0.888
|
247 |
+
- **Macro F1 Score:** 0.941
|
248 |
+
- **Cohen's Kappa:** 0.000
|
249 |
+
- **Matthews Corrcoef:** 0.000
|
250 |
+
|
251 |
+
##### Confusion Matrix
|
252 |
+
|
253 |
+
```
|
254 |
+
Predicted Label 0
|
255 |
+
True Label
|
256 |
+
0 14636
|
257 |
+
```
|
258 |
+
|
259 |
+
##### Classification Report
|
260 |
+
|
261 |
+
```
|
262 |
+
precision recall f1-score support
|
263 |
+
0 1.0 0.888107 0.940738 16480.0
|
264 |
+
micro avg 1.0 0.888107 0.940738 16480.0
|
265 |
+
macro avg 1.0 0.888107 0.940738 16480.0
|
266 |
+
weighted avg 1.0 0.888107 0.940738 16480.0
|
267 |
+
```
|
268 |
+
|
269 |
+
#### IoU Threshold: iou_0.90
|
270 |
+
|
271 |
+
- **Accuracy:** 0.845
|
272 |
+
- **Balanced Accuracy:** 0.845
|
273 |
+
- **Macro Precision:** 1.000
|
274 |
+
- **Macro Recall:** 0.845
|
275 |
+
- **Macro F1 Score:** 0.916
|
276 |
+
- **Cohen's Kappa:** 0.000
|
277 |
+
- **Matthews Corrcoef:** 0.000
|
278 |
+
|
279 |
+
##### Confusion Matrix
|
280 |
+
|
281 |
+
```
|
282 |
+
Predicted Label 0
|
283 |
+
True Label
|
284 |
+
0 13919
|
285 |
+
```
|
286 |
+
|
287 |
+
##### Classification Report
|
288 |
+
|
289 |
+
```
|
290 |
+
precision recall f1-score support
|
291 |
+
0 1.0 0.8446 0.915754 16480.0
|
292 |
+
micro avg 1.0 0.8446 0.915754 16480.0
|
293 |
+
macro avg 1.0 0.8446 0.915754 16480.0
|
294 |
+
weighted avg 1.0 0.8446 0.915754 16480.0
|
295 |
+
```
|
296 |
+
|
297 |
+
#### IoU Threshold: iou_0.95
|
298 |
+
|
299 |
+
- **Accuracy:** 0.758
|
300 |
+
- **Balanced Accuracy:** 0.758
|
301 |
+
- **Macro Precision:** 1.000
|
302 |
+
- **Macro Recall:** 0.758
|
303 |
+
- **Macro F1 Score:** 0.862
|
304 |
+
- **Cohen's Kappa:** 0.000
|
305 |
+
- **Matthews Corrcoef:** 0.000
|
306 |
+
|
307 |
+
##### Confusion Matrix
|
308 |
+
|
309 |
+
```
|
310 |
+
Predicted Label 0
|
311 |
+
True Label
|
312 |
+
0 12490
|
313 |
+
```
|
314 |
+
|
315 |
+
##### Classification Report
|
316 |
+
|
317 |
+
```
|
318 |
+
precision recall f1-score support
|
319 |
+
0 1.0 0.757888 0.862271 16480.0
|
320 |
+
micro avg 1.0 0.757888 0.862271 16480.0
|
321 |
+
macro avg 1.0 0.757888 0.862271 16480.0
|
322 |
+
weighted avg 1.0 0.757888 0.862271 16480.0
|
323 |
+
```
|
324 |
+
|
325 |
+
## Embedding Quality
|
326 |
+
|
327 |
+
### Internal Cluster Validation
|
328 |
+
|
329 |
+
| Silhouette_Score | Davies-Bouldin_Index | Calinski-Harabasz_Index |
|
330 |
+
|-------------------:|-----------------------:|--------------------------:|
|
331 |
+
| 0.291593 | 0.561635 | 3330.14 |
|
332 |
+
|
333 |
+
### External Cluster Validation
|
334 |
+
|
335 |
+
| ARI | NMI | Cluster_Purity |
|
336 |
+
|---------:|---------:|-----------------:|
|
337 |
+
| 0.101185 | 0.317547 | 0.24571 |
|
338 |
+
|
339 |
+
### Mantel Correlation
|
340 |
+
|
341 |
+
| r | p_value | n_items |
|
342 |
+
|-----------:|----------:|----------:|
|
343 |
+
| -0.0955094 | 0.527 | 32 |
|
344 |
+
|