All authors
Claude Skills by wenmin-wu
github.com/wenmin-wu535 skills0 installs688 views
- Dicom Orientation DetectionDecodes MRI scan plane (axial, coronal, sagittal) from DICOM ImageOrientationPatient direction cosine vectors.Votes: 0GitHub stars: 61
- Dicom Voi Lut PreprocessingRead DICOM X-ray files with VOI LUT transformation and MONOCHROME1 inversion for correct pixel intensity renderingVotes: 0GitHub stars: 61
- Differentiable Soft F1 LossUse a soft macro-F1 loss `1 − mean(2·tp / (2·tp + fp + fn))` computed from raw sigmoid probabilities (no thresholding) as a direct training objective for multi-label classification, optionally combined with BCE — closes the gap between training surrogate and the F1 metric the leaderboard scoresVotes: 0GitHub stars: 61
- Dot Annotation Blob Diff ExtractionRecover (x, y, class) point labels from color-coded dot-annotation image pairs via absdiff + blackout masking + Laplacian-of-Gaussian blob detection + center-pixel RGB classificationVotes: 0GitHub stars: 61
- Dual View Reshape ForwardReshape dual-view stacked channels into doubled batch dimension for shared backbone, then concatenate with tabular features for classificationVotes: 0GitHub stars: 61
- Efficientdet Headnet SwapLoad EfficientDet pretrained on COCO with the original 90-class head, then swap in a fresh HeadNet with your own num_classes — keeps the BiFPN feature pyramid pretrained and only retrains the classification head, the canonical transfer-learning recipe for the effdet PyTorch portVotes: 0GitHub stars: 61
- Ema Model AveragingTracks an Exponential Moving Average of model weights during training and evaluates both live and EMA models for more stable predictions.Votes: 0GitHub stars: 61
- Embedding Knn RegressionGPU-accelerated k-NN regression on CLIP image embeddings using cosine distance and inverse-distance-power weighting to predict target embedding vectorsVotes: 0GitHub stars: 61
- Epoch Prediction AveragingCollect test predictions each epoch via callback and combine with exponentially increasing weights favoring later epochsVotes: 0GitHub stars: 61
- Exam Level Label Hierarchy AggregationAggregate per-slice predictions into exam-level labels that satisfy a competition's mutual-exclusion hierarchy (positive vs negative vs indeterminate), using a top-down rule cascade — first decide the exam class, then conditionally rescale the dependent labels so the submission stays internally consistentVotes: 0GitHub stars: 61
- Exam Sequence Padded Mask LossPad variable-length per-slice sequences to a fixed batch length, carry a 0/1 mask alongside, and multiply per-slice BCE by the mask before reducing — gives correct per-exam loss with batched training and zero contamination from padding tokensVotes: 0GitHub stars: 61
- Flat Multicondition HeadModels multiple conditions with a single flat output layer of N_labels × N_classes logits, sliced into per-condition softmax at inference.Votes: 0GitHub stars: 61
- Focal LossAlpha-weighted focal loss that down-weights easy examples to focus training on hard, misclassified pixels in imbalanced segmentation tasks.Votes: 0GitHub stars: 61
- Frame Differencing Temporal EncodingEncode motion and velocity by computing per-channel pixel differences between consecutive frames instead of stacking raw frames for RL visual observationsVotes: 0GitHub stars: 61
- Frame Prediction AveragingAverage per-frame sigmoid predictions across sampled video frames to produce a stable video-level classification probabilityVotes: 0GitHub stars: 61
- Frozen Batchnorm FinetuningUnfreezes backbone layers for fine-tuning while keeping BatchNorm layers frozen to preserve pretrained running statistics.Votes: 0GitHub stars: 61
- Gaussian Sphere Target GenerationGenerates 3D segmentation training targets by placing Gaussian spheres at annotated point coordinates.Votes: 0GitHub stars: 61
- Gem PoolingReplaces global average pooling with Generalized Mean (GeM) pooling, using a learnable or fixed exponent to emphasize high-activation regions.Votes: 0GitHub stars: 61
- Generative Output Numeric CleaningClean noisy numeric strings from generative model output by removing invalid characters, fixing malformed floats, and handling multiple decimal pointsVotes: 0GitHub stars: 61
- Gradient AccumulationAccumulates gradients over multiple mini-batches before stepping the optimizer, simulating larger effective batch sizes.Votes: 0GitHub stars: 61
- Greedy Mask Overlap ResolutionResolve overlapping instance masks by greedily assigning contested pixels to higher-confidence predictions using a running occupancy mapVotes: 0GitHub stars: 61
- Hair Overlay AugmentationOverlay real hair PNGs (masked via threshold) onto dermoscopy images to simulate body-hair occlusion as a domain-specific augmentationVotes: 0GitHub stars: 61
- Heavy Augmentation PipelineComprehensive albumentations augmentation combining geometric, photometric, noise, blur, and cutout transforms for robust CV training.Votes: 0GitHub stars: 61
- Histopathology Image InversionInverts whole slide image pixel values (1 - x) so white background becomes zero, enabling standard zero-padding and making tissue regions the active signal.Votes: 0GitHub stars: 61
- Hu WindowingApply radiological windowing to HU images — clamp to center/width range for tissue-specific visualization (lung, bone, soft tissue)Votes: 0GitHub stars: 61
- Hungarian Matching Detection EvalEvaluates 3D object detection by matching predicted and ground-truth coordinates via the Hungarian algorithm, then computing F-beta score.Votes: 0GitHub stars: 61
- Iou Threshold SweepGrid search binarization thresholds on validation predictions to find the cutoff that maximizes mean IoUVotes: 0GitHub stars: 61
- Iou Weighted Assignment MetricEvaluation scorer that merges predictions with GT per frame, takes top-IoU match per GT, and computes weighted accuracy with IoU threshold gateVotes: 0GitHub stars: 61
- Isotropic Resize With PaddingResize images preserving aspect ratio then zero-pad to a square to avoid distortion artifacts in face crops or object detection inputsVotes: 0GitHub stars: 61
- Isotropic Voxel ResamplingResample 3D CT volumes to uniform voxel spacing using scipy zoom, normalizing physical dimensions across scannersVotes: 0GitHub stars: 61
- Keypoint Aware Raster AugmentationUse albumentations keypoint_params to jointly augment BEV rasters and trajectory target points so the spatial transform stays consistentVotes: 0GitHub stars: 61
- Kfold Model AveragingAverage predictions from K independently trained fold models at inference time for variance reduction without stacking complexityVotes: 0GitHub stars: 61
- Kmeans Dominant Color ExtractionExtract an image's dominant RGB color via k-means over pixel-color space and emit three dense features capturing the modal color of the subjectVotes: 0GitHub stars: 61
- Knn Distance Threshold MatchingKNN-based retrieval with grid-searched distance threshold to convert embedding neighbors into match predictionsVotes: 0GitHub stars: 61
- Lap Hard Negative MiningUse linear assignment problem (LAP/lapjv) on a score matrix to select globally optimal hard-negative pairs for metric learningVotes: 0GitHub stars: 61
- Laplacian Variance Blur ScoreScore image sharpness with the variance of the Laplacian (Pech-Pacheco) as a single scalar feature for downstream tabular models or as a hard blur filterVotes: 0GitHub stars: 61
- Lateralized Label Flip Tta DisableDisable horizontal-flip augmentation (both train-time and TTA) when label columns encode left/right anatomy — flipping silently corrupts the targets because "Left ICA" must map to "Right ICA" after a flip, not stay as "Left ICA"Votes: 0GitHub stars: 61
- Learned Distance MetricTrainable nonlinear distance metric that transforms (v1-v2) and (v1-v2)^2 through a linear layer before computing squared normVotes: 0GitHub stars: 61
- Levenshtein Distance MetricEvaluates image-to-sequence models using mean Levenshtein edit distance between predicted and ground-truth strings.Votes: 0GitHub stars: 61
- Local Contrast EnhancementSubtracts a Gaussian-blurred version of the image from itself to normalize local illumination and enhance fine structural details.Votes: 0GitHub stars: 61
- Lovasz Hinge LossLovasz hinge loss that directly optimizes IoU for binary segmentation by computing a convex surrogate via sorted prediction errors and cumulative Jaccard gradients.Votes: 0GitHub stars: 61
- Map Iou Precision SweepCompute mean Average Precision by sweeping IoU thresholds from 0.5 to 0.95 on RLE-encoded instance masks using pycocotoolsVotes: 0GitHub stars: 61
- Mask To Polygon Contour HierarchyConvert binary segmentation masks to Shapely MultiPolygons using cv2 contour hierarchy to correctly handle interior holes, with Douglas-Peucker simplificationVotes: 0GitHub stars: 61
- Metadata Injection BottleneckInject scalar metadata (depth, position, clinical features) into U-Net bottleneck via RepeatVector and Reshape for metadata-aware segmentationVotes: 0GitHub stars: 61
- Microscope Circular Mask AugMask the corners of a dermoscopy image with a random-radius black circle to mimic the dark vignette of a dermatoscope field of viewVotes: 0GitHub stars: 61
- Middle Mip Std Volume ProjectionCompress a 3D medical volume into a 3-channel 2D image by stacking the middle slice, the max-intensity projection across depth, and the per-pixel std across depth — a poor-man's volumetric encoding that lets any pretrained 3-channel 2D CNN ingest a whole series in a single forward passVotes: 0GitHub stars: 61
- Min Area Mask FilteringRemoves predicted segmentation masks below a per-class minimum pixel area threshold to eliminate small false positive regions at inference time.Votes: 0GitHub stars: 61
- Mixed Precision TrainingUses PyTorch AMP autocast and GradScaler for FP16 training, halving memory usage and speeding up training on modern GPUs.Votes: 0GitHub stars: 61
- Mixup Label SmoothingCombines mixup augmentation (linear interpolation of image pairs and their labels) with label smoothing in a single training pipeline for regularization.Votes: 0GitHub stars: 61
- Modality Adaptive Dicom WindowingRoute each DICOM series to a per-modality window-center / window-width pair (CT/CTA/MRA/MRI) before normalization, so the same model can ingest mixed modalities without one modality's intensity range washing out the othersVotes: 0GitHub stars: 61