I have a dataset and I intend to use multi-label learning approach to recognize the various objects present in the images dataset. What is/are the appropriate segmentation approach to use for the multiple objects detection task?
To start with I believe clustering techniques could be of good utility. Please check with K-means clustering approach and then Fuzzy c-means algorithm.
Dear Oladosu Oladimeji
A unique approach for doing multi-label, semi-automated picture segmentation is suggested. Given a limited number of pixels with user-defined labels, one can calculate the likelihood that a random walker starting at each unlabeled pixel would first reach one of the pre-labeled pixels analytically (and rapidly).
When the three strategies are compared in terms of accuracy score, the binary relevance and label powerset techniques will be the most suited for multi-label classification due to their greater accuracy score. This article demonstrated how to construct a multi-label text classification model using the issue transformation approach.
For Multi-Class Classification, use One-Vs-Rest. One-vs-rest (OvR) is a heuristic approach for employing binary classification algorithms for multi-class classification. It is also known as One-vs-All (OvA). It entails dividing the multi-class dataset into a number of binary classification issues.
Kind Regards
Qamar Ul Islam
Dear Oladosu Oladimeji ,
Multi-label classification involves predicting zero or more class labels. Unlike normal classification tasks where class labels are mutually exclusive, multi-label classification requires specialized machine learning algorithms that support predicting multiple mutually non-exclusive classes or “labels.”
https://machinelearningmastery.com/multi-label-classification-with-deep-learning/#:~:text=Multi-label%20classification%20involves%20predicting,exclusive%20classes%20or%20“labels.”
Regards,
Shafagat
Dear Oladosu Oladimeji
Please take a look at the following paper:
Chapter Deep Neural Networks for Biomedical Image Segmentation: Tren...
The article is focused on segmentation of medical images, perhaps one of the architectures mentioned in Table 3 could help you.
Kind regards,
Cecilia-Irene Loeza-Mejía
- Badges
- Science topic