Image steganography is used to hide a secret image inside a cover image in plain sight. Traditionally, the secret data is converted into binary bits and the cover image is manipulated statistically to embed the secret binary bits. Overloading the cover image may lead to distortions and the secret information may become visible. Hence the hiding capacity of the traditional methods are limited. In this paper, a light-weight yet simple deep convolutional autoencoder architecture is proposed to embed a secret image inside a cover image as well as to extract the embedded secret image from the stego image. The proposed method is evaluated using three datasets - COCO, CelebA and ImageNet. Peak Signal-to-Noise Ratio, hiding capacity and imperceptibility results on the test set are used to measure the performance. The proposed method has been evaluated using various images including Lena, airplane, baboon and peppers and compared against other traditional image steganography methods. The experimental results have demonstrated that the proposed method has higher hiding capacity, security and robustness, and imperceptibility performances than other deep learning image steganography methods.

Real face recognition is a challenging problem especially when face images are subject to distortions. This paper presents an approach to tackle partial occlusion distortions present in real face recognition using a single training sample per person. First, original images are partitioned into multiple blocks and Local Binary Patterns are applied as a local descriptor on each block separately. Then, a dimensionality reduction of the resulting descriptors is carried out using Kernel Principle Component Analysis. Once done, a random sampling method is used to select patches at random and hence build several sub-SVM classifiers. Finally, the results from each sub-classifier are combined in order to increase the recognition performance. To demonstrate the usefulness of the approach, experiments were carried on the AR Face Database and obtained results have shown the effectiveness of our technique.

Emotion recognition from facial expressions is a crucial task in affective computing with applications in human-computer interaction, security and healthcare. However, real-world scenarios often involve partial occlusions of facial features due to factors such as headgear, accessories, face masks and environmental obstructions. This paper discusses the effect of synthetic occlusion on facial emotion recognition on images captured in the wild targeting three main similar emotions: surprise, fear and neutral expressions. An analysis on the effect of such occlusion on each emotion is also performed including a discussion on which parts of the face most affect emotion recognition. This paper finally discusses associated challenges and insights in this field.

While there has been a massive increase in research into face recognition, it remains a challenging problem due to conditions present in real life. This paper focuses on the inherently present issue of partial occlusion distortions in real face recognition applications. We propose an approach to tackle this problem. First, face images are divided into multiple patches before local descriptors of Local Binary Patterns and Histograms of Oriented Gradients are applied on each patch. Next, the resulting histograms are concatenated, and their dimensionality is then reduced using Kernel Principle Component Analysis. Once completed, patches are randomly selected using the concept of random sampling to finally construct several sub-Support Vector Machine classifiers. The results obtained from these sub-classifiers are combined to generate the final recognition outcome. Experimental results based on the AR face database and the Extended Yale B database show the effectiveness of our proposed technique.