Hugging Face's logo

OceanCV
/
European_Green_Crab_Classifier

Image Classification
English
European Green Crab
Crab
Citizen Science
Model card Files
xet
Community
ryanpluv commited on
Commit
4fff905
·
verified ·
1 Parent(s): 1062728

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +78 -26
README.md CHANGED
@@ -2,6 +2,7 @@
2
  license: cc-by-nc-2.0
3
  ---
4
 
 
5
  # European Green Crab Classifier
6
  **By Ryan Luvera**
7
 
@@ -25,46 +26,97 @@ This meant the removal of pictures that included only the claw, were too blurry
25
  ### Classes included:
26
  European green crabs and 9 common crab species found in the Salish Sea.
27
 
28
- - Carcinus maenas
29
- - Cancer_productus
30
- - Eriocheir_sinensis
31
- - Hemigrapsus_nudus
32
- - Hemigrapsus_oregonensis
33
- - Metacarcinus_gracilis
34
- - Metacarcinus_magister
35
- - Pugettia_producta
36
- - Romaleon_antennarium
37
- - Telmessus_cheiragonus
38
 
39
- ### Metadata:
 
 
 
 
 
40
 
41
 
42
- ---
43
- # Model Selection
44
- ### YOLOv11 Object Detection Model
45
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
46
 
47
- ### Additional Features: Ultralytics Objects Counting in Regions & Heatmap
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
48
 
49
  ---
50
  # Model Assessment
51
- ### Here are the metrics I used to assess the accuracy and performance of my model during training.
52
 
53
  ### Confusion Matrix
54
-
55
  ### F1 Score
 
56
 
57
- ### Object Detection Model Output
 
 
 
 
 
58
 
59
  ---
60
  # Model Use-case
61
- **Example Proposal:**
62
-
63
- **Example Hypothesis:**
64
- **Example Impact:**
65
-
66
- **Model Justification:**
67
-
68
  ---
69
  # Disclaimer
70
- Credit for all videos used in the dataset go to
 
 
 
2
  license: cc-by-nc-2.0
3
  ---
4
 
5
+
6
  # European Green Crab Classifier
7
  **By Ryan Luvera**
8
 
 
26
  ### Classes included:
27
  European green crabs and 9 common crab species found in the Salish Sea.
28
 
 
 
 
 
 
 
 
 
 
 
29
 
30
+ | Carcinus maenas |
31
+ | :-----------------------------------------------------------------------------------------------------------------------------------------------------------------------: |
32
+ | <img src="https://huggingface.co/OceanCV/European_Green_Crab_Classifier/resolve/main/example_images/Carcinus_maenas.jpg" alt="Carcinus maenas " width="400" height="300"> |
33
+ | "European green crab" |
34
+ | 4044 images |
35
+ | [Citation](https://huggingface.co/OceanCV/European_Green_Crab_Classifier/resolve/main/citations/Carcinus_maenas_citations.txt) |
36
 
37
 
 
 
 
38
 
39
+ | Cancer productus | Eriocheir sinensis | Hemigrapsus nudus |
40
+ | :------------------------------------------------------------------------------------------------------------------------------------------------------------------------: | :----------------------------------------------------------------------------------------------------------------------------------------------------------------------------: | :--------------------------------------------------------------------------------------------------------------------------------------------------------------------------: |
41
+ | <img src="https://huggingface.co/OceanCV/European_Green_Crab_Classifier/resolve/main/example_images/Cancer_productus.jpg" alt="Cancer productus" width="350" height="225"> | <img src="https://huggingface.co/OceanCV/European_Green_Crab_Classifier/resolve/main/example_images/Eriocheir_sinensis.jpg" alt="Eriocheir sinensis" width="350" height="250"> | <img src="https://huggingface.co/OceanCV/European_Green_Crab_Classifier/resolve/main/example_images/Hemigrapsus_nudus.jpg" alt="Hemigrapsus nudus" width="350" height="225"> |
42
+ | "Red rock crab" | "Chinese mitten crab" | "Purple shore crab" |
43
+ | 884 images | 451 images | 1537 images |
44
+ | [Citation](https://huggingface.co/OceanCV/European_Green_Crab_Classifier/resolve/main/citations/Cancer_productus_citations.txt) | [Citation](https://huggingface.co/OceanCV/European_Green_Crab_Classifier/resolve/main/citations/Eriocheir_sinensis_citations.txt) | [Citation](https://huggingface.co/OceanCV/European_Green_Crab_Classifier/resolve/main/citations/Hemigrapsus_nudus_citations.txt) |
45
+
46
+
47
+
48
+ | Hemigrapsus oregonensis | Metacarcinus gracilis | Metacarcinus magister |
49
+ | :--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------: | :----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------: | :-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------: |
50
+ | <img src="https://huggingface.co/OceanCV/European_Green_Crab_Classifier/resolve/main/example_images/Hemigrapsus_oregonensis.jpg" alt="Hemigrapsus oregonensis" width="350" height="225"> | <img src="https://huggingface.co/OceanCV/European_Green_Crab_Classifier/resolve/main/example_images/Metacarcinus_gracilis.jpg" alt="Metacarcinus gracilis" width="350" height="225"> | <img src="https://huggingface.co/OceanCV/European_Green_Crab_Classifier/resolve/main/example_images/Metacarcinus%20magister.jpg" alt=" Metacarcinus magister " width="350" height="225"> |
51
+ | "Yellow shore crab" | "Graceful rock crab" | "Dungeness Crab" |
52
+ | 1245 images | 177 images | 1460 images |
53
+ | [Citation](https://huggingface.co/OceanCV/European_Green_Crab_Classifier/resolve/main/citations/Hemigrapsus_oregonensis_citations.txt) | [Citation](https://huggingface.co/OceanCV/European_Green_Crab_Classifier/resolve/main/citations/Metacarcinus_gracilis_citations.txt) | [Citation](https://huggingface.co/OceanCV/European_Green_Crab_Classifier/resolve/main/citations/Metacarcinus_magister_citations.txt) |
54
+ | | | |
55
+
56
+
57
+ | Pugettia producta | Romaleon antennarium | Telmessus cheiragonus |
58
+ | :--------------------------------------------------------------------------------------------------------------------------------------------------------------------------: | :-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------: | :------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------: |
59
+ | <img src="https://huggingface.co/OceanCV/European_Green_Crab_Classifier/resolve/main/example_images/Pugettia_producta.jpg" alt="Pugettia producta" width="350" height="225"> | <img src="https://huggingface.co/OceanCV/European_Green_Crab_Classifier/resolve/main/example_images/Romaleon%20antennarium.jpg" alt="Romaleon antennarium " width="350" height="225"> | <img src="https://huggingface.co/OceanCV/European_Green_Crab_Classifier/resolve/main/example_images/Telmessus_cheiragonus.jpg" alt=" Telmessus cheiragonus " width="350" height="200"> |
60
+ | "Kelp crab" | "Pacific rock crab" | "Helmet crab" |
61
+ | 826 images | 1640 images | 897 images |
62
+ | [Citation](https://huggingface.co/OceanCV/European_Green_Crab_Classifier/resolve/main/citations/Pugettia_producta_citations.txt) | [Citation](https://huggingface.co/OceanCV/European_Green_Crab_Classifier/resolve/main/citations/Romaleon_antennarium_citations.txt) | [Citation](https://huggingface.co/OceanCV/European_Green_Crab_Classifier/resolve/main/citations/Telmessus%20cheiragonus_citations.txt) |
63
 
64
+ ### Metadata
65
+ Each dataset has different sources which can be found within European_Green_Crab_Classifier/metadata. All data were gathered from iNaturalist and the GBIF "Research Grade" photos dataset. Citations can be found within European_Green_Crab_Classifier/citations or by clicking the link below any crab above.
66
+
67
+ ---
68
+ # Model Selection
69
+ ### YOLOv11 Classification Model
70
+
71
+ Due to the nature of the data having usually one crab per frame, a YOLOv11n Classification model was trained.
72
+
73
+ ```
74
+ # Train the model
75
+
76
+ results = model.train(data="/content/output",
77
+                       epochs=400,
78
+                       pateience=50,
79
+                       stream_buffer=True,
80
+                       imgsz=640,
81
+                       hsv_h=0.0,
82
+                       degrees=180,
83
+                       hsv_s=0.5,
84
+                       hsv_v=0.25,
85
+                       translate=0.0,
86
+                       scale=0.25,
87
+                       shear = 2,
88
+                       fliplr=0.5,
89
+                       mosaic=0.0,
90
+                       erasing=0.02,
91
+
92
+                       )
93
+ ```
94
+ Full training arguments can be found within European_Green_Crab_Classifier/model
95
 
96
  ---
97
  # Model Assessment
 
98
 
99
  ### Confusion Matrix
100
+ <img src="https://huggingface.co/OceanCV/European_Green_Crab_Classifier/resolve/main/model_metrics/confusion_matrix_normalized_fixed.png" alt=" Confusion matrix" width="700" height="500">
101
  ### F1 Score
102
+ <img src="https://huggingface.co/OceanCV/European_Green_Crab_Classifier/resolve/main/model_metrics/f1_fixed.png" alt="F1 score curve" width="700" height="500">
103
 
104
+ ### Precision and Recall
105
+ <img src="https://huggingface.co/OceanCV/European_Green_Crab_Classifier/resolve/main/model_metrics/pr_fixed.png" alt="Precision and Recall curves" width="700" height="500">
106
+ ### European Green Crab False Negatives
107
+ <img src="https://huggingface.co/OceanCV/European_Green_Crab_Classifier/resolve/main/model_metrics/false_negative_cases.png" alt="European Green Crab False Negatives" width="700" height="400">
108
+ ### Interpretation
109
+ This model exhibits fairly good performance within most classes. The European green crab class resulted in 5% false negatives and Eriocheir sinensis/Hemigrapsus oregonensis being flagged the most for false positives. These results are not super surprising. The poor results from this model are likely due to the data feeding into the model not being refined, and the insufficient training epochs. 20 epochs is much too low, and there are many photos of non-crab species littered throughout the data (as shown above). With the next version of this model, I plan to refine the dataset and create a more refined model that is able to be used with more confidence.
110
 
111
  ---
112
  # Model Use-case
113
+ **Example Proposal:** Most European green crab invasions are discovered incidentally. Whether it be by citizen scientists, the general public, or trained survey scientists, more standardized methods of survey for green crabs are needed to fully understand dispersal. Remote beaches on invasion fronts of green crab dispersal should be monitored for green crab molt presence. High-costs for survey effort can be mitigated through drone surveys paired with computer vision models. This model is a perfect contender for drones conducting beach crab molt surveys. If green crabs are found, managers can implement more trapping effort to enhance green crab population suppression.
114
+ - Study area: Remote beaches in Washington, Canada, and Alaska
115
+ - Hypothesis
116
+ - Ho: EGC invasion sites within the fringes of invasion are well accounted for.
117
+ - Ha : EGC molts are more common on remote beaches where trapping isn’t happening heavily.
 
 
118
  ---
119
  # Disclaimer
120
+ **This model and all data it was trained on are licensed under CC BY-NC 4.0.**
121
+ All data were sourced from iNaturalist and the GBIF. Citations for each of the datasets can be found by clicking on the link below each photo in the above dataset section, or by navitagting to European_Green_Crab_Classifier/citations.
122
+