Hello for every body,
Do you know which the best combinaison of features used in the analysis of skin image in the goal to segment it into lesion and background.
Thank you for your help.
Dear Soumaya,
- It highly depends on your datasets, post-processing algorithms, and your analytical urgency, in terms of normal/detected (0, 1), classification (e.g. multiple groups), or identification of specific abnormalities.
- For detection or classification, you may try edge extraction.
- For identification, you may need to design rather sophisticated case-based algorithms.
It appears to me, perhaps, classification may be of your priority and interest.
Best,
J.
http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.301.927&rep=rep1&type=pdf
https://arxiv.org/pdf/1609.03277.pdf
Article Border Detection on Digitized Skin Tumor Images
Dear Rafiee,
Thank you for your response but what I'am going to do is to detect lesion with unsupervised segmentation method (the method that I will use is clear to me, I will use clustering methods) but my problem is how to caracterize the skin image with features and then segment it.
Hello .
A good skin detector that is capable of capturing skin tones under different conditions is important for human-machine interaction applications. In a general situation, skin detectors, such as skin probability maps or Gaussian mixture models, achieve acceptable skin segmentation results. However, the false positive rate increases significantly when the skin tones are in shadow or when skin-like background objects are under similar illumination. In this paper, we propose a novel skin feature learning algorithm based on stacked autoencoders, which are deep neural networks. To overcome the problems encountered in skin segmentation that are caused by different ethnicities and varying illumination conditions, the stacked autoencoders are utilized to learn more discriminative representations of the skin area in both the RGB color space and the HSV color space. Unlike traditional machine learning methods, instead of predicting each pixel individually, our algorithm utilizes blocks to learn the representations and detect the skin areas. The algorithm exploits the learning ability of deep neural networks to learn high-level representations of skin tones. Experiments on test images show that the proposed algorithm achieves acceptable results on several publicly available data sets. To reduce the difficulty of detecting skin pixels in these data sets, the ground truths of these data sets are commonly focused on foreground skin area detection. Our skin detector is also able to detect background areas.
Dear Soumaya,
- All you need to find out is that, at what times what frequencies behave differently in your datasets.
- I would pass the images through time-frequency transformations, such as, wavelet, Hilbert Huang, or even Gabor. Then, you apply clustering methods on decomposed objects. For instance, Wavelet would simply does segmentations and time-frequency component decompositions on your images, at the same time.
- You may check out pages 54-55, 79, 82, 112 of this thesis:
https://scholarlyrepository.miami.edu/cgi/viewcontent.cgi?article=2434&context=oa_dissertations
and page 40 of this paper:
https://pdfs.semanticscholar.org/bb82/9bb10c34ab39d3e873272cc5ad0ad45b32cd.pdf
- Another method would be design and integration of Latent dirichlet allocation.
- If you have a massive datasets, I would run it through deep learning algorithms, which is highly trending.
Best,
J.
- Badges
- Science topic
- Similar topics
- Computer Science
- Images