Deaf people use sign languages to communicate with other people in the community. Although the sign language is known to hearing-impaired people due to its widespread use among them, it is not known much by other people. In this article, we have developed a real-time sign language recognition system for people who do not know sign language to communicate easily with hearing-impaired people. The sign language used in this paper is American sign language. In this study, the convolutional neural network was trained by using dataset collected in 2011 by Massey University, Institute of Information and Mathematical Sciences, and 100% test accuracy was obtained. After network training is completed, the network model and network weights are recorded for the real-time system. In the real-time system, the skin color is determined for a certain frame for hand use, and the hand gesture is determined using the convex hull algorithm, and the hand gesture is defined in real-time using the registered neural network model and network weights. The accuracy of the real-time system is 98.05%.

Breast cancer is one of the most common types of cancer that affects more than 15% of women throughout their lives. Early detection of this type of cancer is crucial for the positive outcome of the treatment process. There are several screening methods for the detection of breast cancer. In this study, a new method based on thermal imaging is proposed for the detection of breast cancer. Four deep learning networks that were successful in object recognition competitions in recent years were used for breast cancer detection from thermographic breast images and as a result of experimental studies, ResNet50 network achieved the highest test performance in detection of breast cancer by 88.89%. This study demonstrated the applicability of deep learning networks in breast cancer detection.

In this work, we focused on liveness detection for facial recognition system's spoofing via fake face movement. We have developed a pupil direction observing system for anti-spoofing in face recognition systems using a basic hardware equipment. Firstly, eye area is being extracted from real time camera by using Haar-Cascade Classifier with specially trained classifier for eye region detection. Feature points have extracted and traced for minimizing person's head movements and getting stable eye region by using Kanade-Lucas-Tomasi (KLT) algorithm. Eye area is being cropped from real time camera frame and rotated for a stable eye area. Pupils are extracted from eye area by using a new improved algorithm subsequently. After a few stable number of frames that has pupils, proposed spoofing algorithm selects a random direction and sends a signal to Arduino to activate that selected direction's LED on a square frame that has totally eight LEDs for each direction. After chosen LED has been activated, eye direction is observed whether pupil direction and LED's position matches. If the compliance requirement is satisfied, algorithm returns data that contains liveness information. Complete algorithm for liveness detection using pupil tracking is tested on volunteers and algorithm achieved high success ratio.