Alisher Abdulkhaev

Alisher Abdulkhaev

ML Engineer @ Browzzin Board Director @ MLTokyo Ph.D. candidate @ University of Tsukuba

Sitemap

A list of all the posts and pages found on the site. For you robots out there is an XML version available for digesting as well.

Pages

Posts

Rebooting AI

less than 1 minute read

Published:

Rebooting AI

portfolio

publications

A New Approach For Video Watermarking Using Genetic Algorithm

Published in EEMKON2015, 2015

In this paper we have proposed a new approach for video watermarking using genetic algorithm.

Recommended citation: A. Abdulkhaev, S.K. Kayhan (2015). "A New Approach For Video Watermarking Using Genetic Algorithm." EEMKON2015

Masters Thesis: A new approach for video watermarking

Published in Thesis, 2016

Digital Watermarking plays a significant role in data authentication and proof of ownership. In this study, a novel approach for Video Watermarking is proposed, in which the Genetic Algorithm is used to optimize the parameters of digital video watermarking scheme. One of the most straightforward watermarking method is LSB (Least Signifiant Bit) in which watermark is embedded into least significant bit of host media. Our work is based on LSB watermarking method. Although, standard LSB watermarking is good choice for data embedding, it is not robust against bit-deletion, gaussian, median filtering attacks. We use genetic algorithm to select the most optimum bit plane during the embedding process and the watermark is embedded into optimum (6th,7th or 8th) bit of host video frames, to overcome above mentioned attacks’ degradations. Decision is done with respect to Peak Signal Noise Ratio (PSNR) of watermarked frame and Normalized Cross Correlation (NCC) values of the original and extracted watermark. Experiments are performed for various watermarks, host video, and a set of Genetic Algorithm parameters, such as population size, number of iteration, tournament range. Results suggest that modified LSB watermarking scheme is more robust to bit-plane deletion, gaussian, median filtering and other similar attacks.

Recommended citation: A. Abdulkhaev (2016). "A new approach for video watermarking" https://alisher-ai.github.io/files/Masters-thesis--A-new-approach-for-video-watermarking.pdf

talks

GAN Workshop

Published:

Machine Learning Tokyo organized a hands-on workshop on GANs (Generative Adversarial Networks) for the Deep Learning community in Tokyo. The session was dynamic and interactive working session, with 3 essential parts:

Object Detection Workshop

Published:

Machine Learning Tokyo organized a workshop on Object Detection for the Deep Learning community in Tokyo. We have covered R-CNN, Fast R-CNN, Faster R-CNN and used the TensorFlow Object Detection API for the implementation part.

Convolutional Operations Workshop

Published:

Machine Learning Tokyo is organized a new deep learning workshop, an Introduction to Convolution Operation @Rakuten. We have covered learn theoretical concepts underlying convolutional neural networks as well as some of the most important CNN architectures, including small interactive implementation blocks.

One-shot learning

Published:

In this talk I have presented about one-shot learning - metric learning, siamese networks - and we had a hands-on implementation of siamese network training.

Learning in Deep Networks

Published:

Machine Learning Tokyo is organized a new deep learning workshop, Part II of our Deep Learning series. In this workshop we have looked at the learning process in Deep Networks, including small interactive implementation blocks.

Convolutional Operations Workshop

Published:

The MLT workshop on CNN Architectures has conducted by our MLT Core Team Engineers Dimitris Katsios, Mustafa Yagmur and Alisher Abdulkhaev.

Data visualization

Published:

The Earth-Life Science Institute at the Tokyo Institute of Technology held a Machine Learning and Deep Learning Bootcamp in collaboration with Machine Learning Tokyo for scientists in Biology, Chemistry, Astro- and Geophysics, and other fields researching the origins of life. Find the MLT materials on GitHub:

MLT Community Hangout

Published:

Over the past 3 years, MLT has evolved from a 2-people study group to one of the biggest and most active technical Machine Learning communities in the world with around 8,000 members. As a nonprofit organization, MLT has tirelessly supported open education, open-source, and open science efforts in Artificial Intelligence in Japan. Over the past year more and more members, team leads, and partners from all over Asia, Europe, and America have joined MLT, and with increasing online efforts, MLT has become an organization that is based in Tokyo but operating globally.

Rethinking Object Detection

Published:

Object localization/detection is one of the most crucial tasks in artificial intelligence and computer vision. The importance of object detection is that most of the vision tasks start with object localization. Real-world applications such as autonomous driving, personal/industrial robotics, person counting, object tracking, surveillance, OCR (optical character recognition) need to localize the object in the given image or video.

teaching

Teaching experience 1

Undergraduate course, University 1, Department, 2014

This is a description of a teaching experience. You can use markdown like any other post.

Teaching experience 2

Workshop, University 1, Department, 2015

This is a description of a teaching experience. You can use markdown like any other post.