GitHub - escuccim/rsna_pneumonia: ConvNets designed for Kaggle's RSNA Pneumonia Detection Challenge

RSNA Pneumonia Detection Kaggle Challenge

In this repo are notebooks and files from the RSNA Pneumonia Detection challenge. I tried a huge number of architectures and approaches but was only able to score a .176 on the leaderboard.

The best performing model was the Mask_RCNN at size 320x320, although I suspect I could have beaten that score given more time or compute.

Name		Name	Last commit message	Last commit date
Latest commit History 17 Commits
Classifier.ipynb		Classifier.ipynb
README.md		README.md
Untitled.ipynb		Untitled.ipynb
YOLO.ipynb		YOLO.ipynb
best-cnn-segmentation-connected-components-384x384.ipynb		best-cnn-segmentation-connected-components-384x384.ipynb
best_320x320.ipynb		best_320x320.ipynb
best_320x320_colab.ipynb		best_320x320_colab.ipynb
best_320x320_colab_2.ipynb		best_320x320_colab_2.ipynb
best_448x448_colab.ipynb		best_448x448_colab.ipynb
best_best_cnn_segmentation_connected_components_384x384_3.ipynb		best_best_cnn_segmentation_connected_components_384x384_3.ipynb
best_cnn_segmentation_connected_components_384x384_2.ipynb		best_cnn_segmentation_connected_components_384x384_2.ipynb
best_cnn_segmentation_connected_components_384x384_6.ipynb		best_cnn_segmentation_connected_components_384x384_6.ipynb
best_cnn_segmentation_connected_components_384x384_ff.ipynb		best_cnn_segmentation_connected_components_384x384_ff.ipynb
best_cnn_segmentation_connected_components_512x512_1.ipynb		best_cnn_segmentation_connected_components_512x512_1.ipynb
best_cnn_segmentation_connected_components_512x512_2.ipynb		best_cnn_segmentation_connected_components_512x512_2.ipynb
best_cnn_segmentation_connected_components_512x512_3.ipynb		best_cnn_segmentation_connected_components_512x512_3.ipynb
best_cnn_segmentation_connected_components_512x512_4.ipynb		best_cnn_segmentation_connected_components_512x512_4.ipynb
best_cnn_segmentation_connected_components_512x512_4.py		best_cnn_segmentation_connected_components_512x512_4.py
cnn-segmentation-connected-components.ipynb		cnn-segmentation-connected-components.ipynb
cnn_segmentation_connected_components_1.ipynb		cnn_segmentation_connected_components_1.ipynb
cnn_segmentation_connected_components_2.ipynb		cnn_segmentation_connected_components_2.ipynb
cnn_segmentation_connected_components_3.ipynb		cnn_segmentation_connected_components_3.ipynb
cnn_segmentation_connected_components_4.ipynb		cnn_segmentation_connected_components_4.ipynb
mask-rcnn-and-transfer-learning.ipynb		mask-rcnn-and-transfer-learning.ipynb
mask_rcnn_and_transfer_learning_256x256_2.ipynb		mask_rcnn_and_transfer_learning_256x256_2.ipynb
mask_rcnn_and_transfer_learning_320x320.ipynb		mask_rcnn_and_transfer_learning_320x320.ipynb
mask_rcnn_and_transfer_learning_320x320_colab.ipynb		mask_rcnn_and_transfer_learning_320x320_colab.ipynb
mask_rcnn_and_transfer_learning_384x384.ipynb		mask_rcnn_and_transfer_learning_384x384.ipynb
mask_rcnn_and_transfer_learning_384x384_2.ipynb		mask_rcnn_and_transfer_learning_384x384_2.ipynb
mask_rcnn_and_transfer_learning_448x448-Copy2.ipynb		mask_rcnn_and_transfer_learning_448x448-Copy2.ipynb
mask_rcnn_and_transfer_learning_448x448-Copy3.ipynb		mask_rcnn_and_transfer_learning_448x448-Copy3.ipynb
mask_rcnn_and_transfer_learning_448x448.ipynb		mask_rcnn_and_transfer_learning_448x448.ipynb
mean-average-precision-metric.ipynb		mean-average-precision-metric.ipynb
non_max_suppression.ipynb		non_max_suppression.ipynb
preprocess_yolo_labels.ipynb		preprocess_yolo_labels.ipynb
preprocess_yolo_labels_10.ipynb		preprocess_yolo_labels_10.ipynb
preprocess_yolo_labels_11.ipynb		preprocess_yolo_labels_11.ipynb
preprocess_yolo_labels_2.ipynb		preprocess_yolo_labels_2.ipynb
preprocess_yolo_labels_3.ipynb		preprocess_yolo_labels_3.ipynb
preprocess_yolo_labels_4.ipynb		preprocess_yolo_labels_4.ipynb
preprocess_yolo_labels_5.ipynb		preprocess_yolo_labels_5.ipynb
preprocess_yolo_labels_6.ipynb		preprocess_yolo_labels_6.ipynb
preprocess_yolo_labels_7.ipynb		preprocess_yolo_labels_7.ipynb
preprocess_yolo_labels_8.ipynb		preprocess_yolo_labels_8.ipynb
preprocess_yolo_labels_9.ipynb		preprocess_yolo_labels_9.ipynb
sample_images.ipynb		sample_images.ipynb
yolo-for-detection-of-bounding-boxes-tensorflow.ipynb		yolo-for-detection-of-bounding-boxes-tensorflow.ipynb
yolo_448X448_21_colab.ipynb		yolo_448X448_21_colab.ipynb
yolo_448X448_23_colab.ipynb		yolo_448X448_23_colab.ipynb
yolo_448X448_24.ipynb		yolo_448X448_24.ipynb
yolo_448X448_24_4.ipynb		yolo_448X448_24_4.ipynb
yolo_448X448_25.ipynb		yolo_448X448_25.ipynb
yolo_448X448_25_2.ipynb		yolo_448X448_25_2.ipynb
yolo_448X448_26.ipynb		yolo_448X448_26.ipynb
yolo_448X448_26_colab.ipynb		yolo_448X448_26_colab.ipynb
yolo_448X448_27_1_colab.ipynb		yolo_448X448_27_1_colab.ipynb
yolo_448X448_28_1_colab.ipynb		yolo_448X448_28_1_colab.ipynb
yolo_448X448_29_1_colab.ipynb		yolo_448X448_29_1_colab.ipynb
yolo_448X448_30_1_colab.ipynb		yolo_448X448_30_1_colab.ipynb
yolo_448X448_32_1_colab.ipynb		yolo_448X448_32_1_colab.ipynb
yolo_448X448_34_1_colab.ipynb		yolo_448X448_34_1_colab.ipynb
yolo_448X448_35_1_colab.ipynb		yolo_448X448_35_1_colab.ipynb
yolo_448X448_36_1_colab.ipynb		yolo_448X448_36_1_colab.ipynb
yolo_480x480_15.ipynb		yolo_480x480_15.ipynb
yolo_480x480_16.ipynb		yolo_480x480_16.ipynb
yolo_480x480_18.ipynb		yolo_480x480_18.ipynb
yolo_480x480_18_colab.ipynb		yolo_480x480_18_colab.ipynb
yolo_480x480_19 (1).ipynb		yolo_480x480_19 (1).ipynb
yolo_480x480_19.ipynb		yolo_480x480_19.ipynb
yolo_480x480_20.ipynb		yolo_480x480_20.ipynb
yolo_480x480_20_colab.ipynb		yolo_480x480_20_colab.ipynb
yolo_512x512_1.ipynb		yolo_512x512_1.ipynb
yolo_512x512_10.ipynb		yolo_512x512_10.ipynb
yolo_512x512_11.ipynb		yolo_512x512_11.ipynb
yolo_512x512_12-Copy2.ipynb		yolo_512x512_12-Copy2.ipynb
yolo_512x512_12.ipynb		yolo_512x512_12.ipynb
yolo_512x512_13.ipynb		yolo_512x512_13.ipynb
yolo_512x512_13_colab.ipynb		yolo_512x512_13_colab.ipynb
yolo_512x512_14.ipynb		yolo_512x512_14.ipynb
yolo_512x512_15.ipynb		yolo_512x512_15.ipynb
yolo_512x512_17_colab.ipynb		yolo_512x512_17_colab.ipynb
yolo_512x512_2.ipynb		yolo_512x512_2.ipynb
yolo_512x512_3.ipynb		yolo_512x512_3.ipynb
yolo_512x512_4.ipynb		yolo_512x512_4.ipynb
yolo_512x512_5.ipynb		yolo_512x512_5.ipynb
yolo_512x512_6.ipynb		yolo_512x512_6.ipynb
yolo_512x512_7.ipynb		yolo_512x512_7.ipynb
yolo_512x512_8.ipynb		yolo_512x512_8.ipynb
yolo_512x512_8_working.ipynb		yolo_512x512_8_working.ipynb
yolo_512x512_8_working_better.ipynb		yolo_512x512_8_working_better.ipynb
yolo_512x512_9.ipynb		yolo_512x512_9.ipynb
yolov3-rsna-starting-notebook.ipynb		yolov3-rsna-starting-notebook.ipynb

escuccim/rsna_pneumonia

Folders and files

Latest commit

History