International Skin Imaging Collaboration (ISIC)

Click here to visit the external workshop website

Teaser picture for paper
    Authors: M. Emre Celebi, Noel C. F. Codella, Marc Combalia, Kristin Dana, Allan Halpern, Philipp Tschandl   
Mond Jun15  
7:45 AM - 8:00 AM
Favorite
Teaser picture for paper
Skin appearance modeling using high resolution photography has led to advances in recognition, synthesis and prediction. Computational appearance pro
    Authors: Kristin Dana   
    Keywords:  skin texture,skin recognition,microbiome, multimodal,UV,blue light, fluorescence,deep learning, texture transfer, style transfer
Mond Jun15  
8:00 AM - 8:30 AM
Favorite
Teaser picture for paper
This presentation examines ways to improve the current AI systems for melanoma detection.
    Authors: Tim K Lee   
    Keywords:  melanoma, computer-aided diagnosis, black-box, interpretability, clinical features, dermoscopic features, 7-point checklist, polarization speckle, roughness, polarization
Mond Jun15  
8:30 AM - 9:00 AM
Favorite
Teaser picture for paper
    Authors: Doug Canfield   
Mond Jun15  
9:00 AM - 9:30 AM
Favorite
Teaser picture for paper
Skinvision-21st century medical device melanoma screening At SkinVision, we take the health of our users seriously. Research is at the heart of our o
    Authors: Daniel Mark Siegel   
    Keywords:  Melanoma,Skinvision,deep neural networks,artificial intelligence,augmented intelligence, Skin cancer,squamous cell carcinoma,basal cell carcinoma,teledermatology,dermatology
Mond Jun15  
9:30 AM - 10:00 AM
Favorite
Teaser picture for paper
We incorporate information from illumination and color imaging of the skin to improve deep learning-based skin lesion segmentation performance.
    Authors: Kumar Abhishek, Ghassan Hamarneh, Mark S Drew   
    Keywords:  skin lesion, segmentation, illumination, color imaging, deep learning
Mond Jun15  
10:00 AM - 10:12 AM
Favorite
Teaser picture for paper
Proposed approach, experimental results on the 3 skin lesion datasets, along with comparison between the meta-learning techniques and baseline.
    Authors: Kushagra Mahajan, Monika Sharma, Lovekesh Vig   
    Keywords:  Skin-diseases,Classification,Few-shot,Meta-learning,Reptile,Prototypical-networks,G-convolutions,Rare-diseases
Mond Jun15  
10:12 AM - 10:24 AM
Favorite
Teaser picture for paper
Computer-aided skin cancer detection systems built with deep neural networks yield overconfident predictions on out-of-distribution examples. Motivate
    Authors: Andre GC Pacheco, Chandramouli Sastry, Thomas Trappenberg, Sageev Oore, Renato Krohling   
    Keywords:  Out-of-distribution detection, deep neural networks, skin cancer detection, gram matrix, classification
Mond Jun15  
10:24 AM - 10:36 AM
Favorite
Teaser picture for paper
A MTL framework based on "gates" that learn what to share between tasks. Gates can be examined to understand the relationships learned by the model.
    Authors: Davide Coppola, Hwee Kuan Lee, Cuntai Guan   
    Keywords:  7-point checklist, multi-task learning, explainable ai, melanoma, gates, feature-sharing, cnn, xai, mtl, deep learning
Mond Jun15  
10:36 AM - 10:48 AM
Favorite
Teaser picture for paper
A comparison between skin disease classification model explanations and human-identified regions of interest.
    Authors: Nalini M Singh, Kang Lee, David Coz, Christof Angermueller, Susan Huang, Aaron Loh, Yuan Liu   
    Keywords:  interpretability, model explanation, saliency maps
Mond Jun15  
10:48 AM - 11:00 AM
Favorite
Teaser picture for paper
We improve skin segmentation task with less data, by selecting samples with the best Cohen’s Kappa, and conditioning the masks to remove detail.
    Authors: Vinicius Ribeiro   
    Keywords:  skin lesion segmentation; inter-annotator agreement; deep learning; sample selection; label conditioning; less is more; cohen's kappa; skin cancer
Mond Jun15  
11:00 AM - 11:12 AM
Favorite
Teaser picture for paper
We address skin-lesion dataset bias with two objectives: finding what are the spurious correlations exploited by biased networks, and unbiasing them.
    Authors: Alceu E Bissoto, Eduardo Valle, Sandra Avila    
    Keywords:  skin cancer,bias,debias,generalization,feature analysis,classification,interpretability,deep learning
Mond Jun15  
11:12 AM - 11:24 AM
Favorite
Teaser picture for paper
How to make the most of the available dermoscopy data? This paper answers this question comparing several sample weighting methods.
    Authors: Catarina Barata, Carlos Santiago   
    Keywords:  Imbalanced datasets, sample weighting, class-balancing losses, curriculum learning, deep learning, skin cancer
Mond Jun15  
11:24 AM - 11:36 AM
Favorite
Teaser picture for paper
We explore uncertainty estimation techniques and metrics based on Monte-Carlo sampling and apply them to the problem of skin lesion classification.
    Authors: Marc Combalia, Ferran Hueto, Susana Puig, Josep Malvehy, Veronica Vilaplana   
    Keywords:  uncertainty, confidence, healthcare, skin, classification, deep learning, neural network
Mond Jun15  
11:36 AM - 11:48 AM
Favorite
Teaser picture for paper
Our work explores stacked ensemble learning for Skin Lesion classification, augmented for novelty detection via the proposed CS-KSU modules.
    Authors: Subhranil Bagchi, Anurag Banerjee, Deepti Bathula   
    Keywords:  skin cancer, stacking, meta learning, novelty detection, unknown class, data imbalance, classification, hierarchical model
Mond Jun15  
11:48 AM - 12:00 PM
Favorite
Teaser picture for paper
    Authors: M. Emre Celebi, Noel C. F. Codella, Marc Combalia, Kristin Dana, Allan Halpern, Philipp Tschandl   
Mond Jun15  
12:00 PM - 12:30 PM
Favorite
Teaser picture for paper
    Authors: M. Emre Celebi, Noel C. F. Codella, Marc Combalia, Kristin Dana, Allan Halpern, Philipp Tschandl   
Mond Jun15  
12:30 PM - 1:00 PM
Favorite