Hugging Face's logo

Reduanul1997
/
RepositoryBaseLine3DuNet

Model card Files
xet
 Community
RepositoryBaseLine3DuNet
File size: 3,836 Bytes
acb09cb
 
 
5973791
acb09cb
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
5973791
 
 
 
 
 
 
 
 
 
 
acb09cb
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
{
    "archive_metadata": {
        "archive_name": "Baseline 3D U-Net Final Evaluation Reference",
        "created_utc": "2026-09-02T17:50:33.575997+00:00",
        "evaluation_block": "Block 13",
        "archive_block": "Block 14",
        "status": "FINAL_BASELINE_REFERENCE"
    },
    "repository": {
        "huggingface_repo": "Reduanul1997/RepositoryBaseLine3DuNet",
        "repo_type": "model"
    },
    "model": {
        "name": "Original3DUNet",
        "architecture": "Baseline 3D U-Net",
        "input_channels": 4,
        "output_classes": 4,
        "parameters": 22580964
    },
    "dataset": {
        "dataset": "BraTS2020",
        "modalities": [
            "FLAIR",
            "T1",
            "T1ce",
            "T2"
        ],
        "classes": [
            "Background",
            "NCR/NET",
            "Edema",
            "Enhancing Tumor"
        ],
        "training_patients": 294,
        "validation_patients": 74,
        "patch_size": [
            64,
            64,
            64
        ]
    },
    "model_selection": {
        "best_model_epoch": 45,
        "best_validation_dice_from_training": 0.772125,
        "note": "This is the validation Dice reported by the training pipeline and used for model selection. It is distinct from the independently recomputed full-validation confusion-matrix metrics."
    },
    "evaluation_reference": {
        "foreground_macro": {
            "dice": 0.757824,
            "iou": 0.618727,
            "precision": 0.781685,
            "recall": 0.740964
        },
        "tumor_binary": {
            "dice": 0.927715,
            "iou": 0.865175,
            "precision": 0.942361,
            "recall": 0.913516,
            "sensitivity": 0.913516,
            "false_negative_rate": 0.086484,
            "false_positive_rate": 0.017968,
            "TP": 4312069,
            "TN": 14414613,
            "FP": 263746,
            "FN": 408228
        },
        "per_class": {
            "Unknown": {
                "Model": "Original3DUNet",
                "Best_Model_Epoch": 45.0,
                "Class_ID": 3.0,
                "Class_Name": "Enhancing Tumor",
                "Dice": 0.8352533422346283,
                "IoU": 0.7171116024810977,
                "Precision": 0.8498932748404121,
                "Recall": 0.8211092324031519,
                "Sensitivity": 0.8211092324031519,
                "FNR": 0.1788907675968382,
                "FPR": 0.0080306550029849,
                "TP": 835757.0,
                "TN": 18233207.0,
                "FP": 147610.0,
                "FN": 182082.0
            }
        }
    },
    "baseline_comparison_policy": {
        "primary_segmentation_metric": "Foreground Macro Dice",
        "secondary_segmentation_metrics": [
            "Foreground Macro IoU",
            "Foreground Macro Precision",
            "Foreground Macro Recall"
        ],
        "primary_fn_metric": "Tumor False Negative Rate (FNR)",
        "additional_fn_metric": "Tumor False Negatives (FN)",
        "additional_tumor_metrics": [
            "Tumor Dice",
            "Tumor IoU",
            "Tumor Precision",
            "Tumor Recall",
            "Tumor Sensitivity",
            "Tumor FPR"
        ],
        "comparison_rule": "Future models must be evaluated using the same validation split, preprocessing, class definitions, tumor-vs-background definition, and metric calculation procedure before direct comparison with this baseline."
    },
    "artifact_files": [
        "baseline_evaluation_results.json",
        "baseline_evaluation_metrics.csv",
        "baseline_per_class_metrics.csv",
        "baseline_multiclass_confusion_matrix.png",
        "baseline_tumor_background_confusion_matrix.png",
        "baseline_final_evaluation_summary.json"
    ]
}