All authors
Claude Skills by wenmin-wu
github.com/wenmin-wu535 skills0 installs688 views
- Morphological Lung SegmentationSegment lung regions from CT using HU thresholding, connected-component labeling, and morphological openingVotes: 0GitHub stars: 61
- Multi Annotator Bbox ConsensusMerges overlapping same-class bounding boxes from multiple annotators into a consensus box using IoU-based matching and intersection.Votes: 0GitHub stars: 61
- Multi Efficientnet Shared InputCombine EfficientNetB0..B6 into one Keras model with a shared image input and one sigmoid head per backbone, training all N models in a single fit() call on TPUVotes: 0GitHub stars: 61
- Multi Scale Patch Training PyramidGenerate count-regression training patches at a geometric pyramid of image scales (0.9^k) so one CNN handles within- and between-image object-size variation without explicit anchorsVotes: 0GitHub stars: 61
- Multi Series Channel StackingStacks uniformly sampled slices from multiple MRI series (e.g., Sagittal T1, T2, Axial) into a single multi-channel tensor for one-pass inference.Votes: 0GitHub stars: 61
- Multi Window Channel StackingStack multiple CT window settings (brain, subdural, bone) as separate RGB channels for CNN inputVotes: 0GitHub stars: 61
- Multichannel To Rgb AdaptationComposites multi-channel imagery (microscopy, satellite) into 3-channel RGB for pretrained CNN backbones.Votes: 0GitHub stars: 61
- Multilabel Auc EvaluationComputes per-class ROC-AUC then macro-averages for multi-label classification evaluation and model selection.Votes: 0GitHub stars: 61
- Multilabel Cooccurrence Correlation EdaFor multi-label classification, compute the per-class binary correlation matrix restricted to multi-label rows and the conditional class counts given a rare anchor class — reveals label groupings the model can exploit (shared classifier heads, hierarchical loss weighting, post-hoc consistency rules)Votes: 0GitHub stars: 61
- Multilabel Rare Class Image OversamplingOversample multi-label images by giving each image a duplication multiplier equal to the max per-class multiplier among its labels, so every rare class gets repetition without exploding common-class counts — the standard fix for long-tail multi-label distributions where SMOTE / per-row oversampling doesn't applyVotes: 0GitHub stars: 61
- Multimodal Prediction UnionCombine match predictions from image embeddings, text similarity, and perceptual hash via set union for maximum recallVotes: 0GitHub stars: 61
- Nested Unet Dense Skip ConnectionsUNet++ dense cross-depth skip connections that propagate deeper decoder features into all shallower decoder levelsVotes: 0GitHub stars: 61
- Numeric Categorical Auto DetectionAuto-detect whether a generated data series is numeric or categorical by measuring the fraction of digit characters in the concatenated valuesVotes: 0GitHub stars: 61
- Open Set Distance CutoffAssign an unknown/novel class when all nearest-neighbor distances exceed a tuned cutoff threshold for open-set recognitionVotes: 0GitHub stars: 61
- Optimized Ordinal ThresholdsUses Nelder-Mead optimization to find per-class decision thresholds that maximize Quadratic Weighted Kappa for regression-to-ordinal conversion.Votes: 0GitHub stars: 61
- Ordinal Multilabel EncodingEncodes ordinal classes as cumulative binary labels (class N activates labels 0..N), enabling sigmoid + BCE training for ordinal regression.Votes: 0GitHub stars: 61
- Pack Padded Sequence LossUses pack_padded_sequence to exclude padding tokens from cross-entropy loss in variable-length sequence generation.Votes: 0GitHub stars: 61
- Pairwise Distance Proximity FilterCompute Euclidean distance between entity pairs from tracking data and filter out pairs beyond a threshold to reduce inference candidatesVotes: 0GitHub stars: 61
- Pairwise Tracking Feature MergeDouble left-join on tracking data to create pairwise features (positions, velocities, distance) for both entities in an interaction pairVotes: 0GitHub stars: 61
- Patch Grid Count RegressionTile a large aerial image into fixed-size patches, accumulate per-class point-annotation counts into a grid tensor aligned with the tiles, and train a small CNN to regress per-class object counts per patch under MSEVotes: 0GitHub stars: 61
- Patient Level Stratified KfoldStratifies CV folds at the patient level rather than image level, preventing data leakage when multiple images exist per patient.Votes: 0GitHub stars: 61
- Per Class Score ThresholdApply class-specific confidence thresholds by inferring the dominant class per image and indexing into a per-class threshold arrayVotes: 0GitHub stars: 61
- Per Class Soft F1 Threshold FittingOptimize per-class decision thresholds for macro-F1 by replacing the non-differentiable hard threshold with a sigmoid-sharpened soft-F1 surrogate and fitting the per-class threshold vector via least-squares — averaged over multiple random validation splits to suppress overfitting on rare classesVotes: 0GitHub stars: 61
- Per Label Platt Isotonic CalibrationFit a per-label probability calibrator on out-of-fold scores using Platt scaling (logistic regression on raw scores) and fall back to isotonic regression for labels where the logistic doesn't converge — pickle the dict of fitted calibrators and apply at inference for a small but free leaderboard lift on multi-label classificationVotes: 0GitHub stars: 61
- Per Modality Ensemble AveragingTrain separate models per imaging modality (FLAIR/T1w/T1wCE/T2w) and average their predictions for final ensembleVotes: 0GitHub stars: 61
- Per Modality Separate ModelTrains one specialized model per imaging modality or series type, routing inputs by metadata at inference for modality-specific feature learning.Votes: 0GitHub stars: 61
- Per Organ Multihead Sigmoid SoftmaxSingle CNN backbone with one shallow Dense neck per organ and mixed sigmoid (binary) + softmax (multi-class severity) heads, trained with a dict of losses so each organ is calibrated independently while sharing visual featuresVotes: 0GitHub stars: 61
- Per Patient Slice Maxpool AggregationAggregate per-slice CNN predictions into a single patient-level injury score by mean-pooling across TTA copies first, then max-pooling across slices — the worst-slice wins, which matches the medical reality that one bad slice is enough to grade the patientVotes: 0GitHub stars: 61
- Percentile Contrast StretchNormalize high-dynamic-range satellite or medical imagery to [0,1] using per-channel percentile clipping to suppress outliers while preserving relative contrastVotes: 0GitHub stars: 61
- Pfbeta Threshold OptimizationGrid-searches the optimal classification threshold to maximize probabilistic F-beta score on validation predictions.Votes: 0GitHub stars: 61
- Phash Duplicate GroupingGroup near-duplicate images by perceptual hash (pHash) as a zero-cost baseline signal for product or image matchingVotes: 0GitHub stars: 61
- Point Centered Fixed Patch CroppingConvert (x, y, class) point annotations into a CNN classification training set by cropping fixed-size square patches centered on each point, using a numpy shape check to silently reject border-clipped cropsVotes: 0GitHub stars: 61
- Prediction Map Stitching AveragingStitch overlapping tile predictions into a full-resolution output by accumulating probabilities and dividing by per-pixel overlap countsVotes: 0GitHub stars: 61
- Progressive Dropout UnetApply lower dropout in shallow/final U-Net layers and higher dropout in deep layers to preserve spatial detail while regularizing abstract featuresVotes: 0GitHub stars: 61
- Quadratic Weighted Kappa CallbackCustom training callback that computes Quadratic Weighted Kappa on validation data each epoch and checkpoints the best model.Votes: 0GitHub stars: 61
- Quantile Slice SamplingSamples a fixed number of slices from variable-length CT/MRI stacks using quantile indexing to produce consistent input depth.Votes: 0GitHub stars: 61
- Quantile Threshold Prevalence MatchingSets the binary classification threshold as a prediction quantile matching the expected positive prevalence rate, avoiding manual threshold tuning.Votes: 0GitHub stars: 61
- Raft Optical Flow ExtractionExtract dense per-pixel motion fields between consecutive video frames using a pretrained RAFT model, producing an HxWx2 flow tensor that can be channel-stacked with RGB or used as a standalone motion feature for action / impact / event detectionVotes: 0GitHub stars: 61
- Raster To Svg Polygon ConversionConverts a raster image to a size-bounded SVG via K-means color quantization, contour extraction, importance-ranked polygon assembly, and progressive simplification.Votes: 0GitHub stars: 61
- Rgby 4channel Fluorescence LoaderLoad fluorescence microscopy images stored as 4 separate single-channel PNGs (red microtubules, green target protein, blue nucleus, yellow ER) into a single HxWx4 tensor, preserving the biological semantics of each channel rather than collapsing to RGBVotes: 0GitHub stars: 61
- Rle Mask EncodingEncodes binary segmentation masks into compressed RLE format for efficient storage and submission.Votes: 0GitHub stars: 61
- Rotation Search Point RegistrationBrute-force a 2D rotation angle over a coarse grid to align field-coordinate points with image-plane detections when the camera angle is unknownVotes: 0GitHub stars: 61
- Rotation Tta SegmentationTest-time augmentation via 4 rotation angles (0/90/180/270), applying inverse rotation to each prediction before averagingVotes: 0GitHub stars: 61
- Semi Supervised Pretrained BackboneUses Facebook's semi-weakly supervised ImageNet-pretrained models (trained on 940M unlabeled images) as CNN backbones for stronger transfer learning than standard supervised pretraining.Votes: 0GitHub stars: 61
- Sentence Transformer Target EncodingEncode text prompts into fixed-length dense vectors using SentenceTransformer for cosine-similarity evaluation in image-to-text retrieval tasksVotes: 0GitHub stars: 61
- Separable Temporal Spectral Cnn2D CNN with asymmetric kernels — temporal convolutions (Nx1) then spectral convolutions (1xM) — to decouple time and feature extractionVotes: 0GitHub stars: 61
- Separate Pos Neg Dice TrackingTrack dice score separately for positive (mask-present) and negative (empty-mask) images to avoid division distortionVotes: 0GitHub stars: 61
- Siamese Pairwise Comparison HeadSiamese network head that compares two embeddings via element-wise multiply, add, abs-diff, and squared-diff features for verification tasksVotes: 0GitHub stars: 61
- Sigmoid Normalized RmseSigmoid-transformed normalized RMSE that maps error from [0,inf) to a bounded (0,1] similarity score using R2-score ratioVotes: 0GitHub stars: 61
- Slice As Channel 2d CnnResample a 3D medical volume to a fixed depth N (e.g. 32) and feed the N slices as input *channels* to a 2D CNN with `in_chans=N` instead of using a 3D conv backbone — gets the volumetric context for a fraction of the memory and lets you use any timm 2D pretrained modelVotes: 0GitHub stars: 61