Every individual's perception of multimedia content varies based on their interpretation. Therefore, it is quite challenging to predict likability of any multimedia just based on its content. This paper presents a novel system for analysis of facial expressions of subject against the multimedia content to be evaluated. First, we developed a dataset by recording facial expressions of subjects under uncontrolled environment. These subjects are volunteers recruited to watch the videos of different genre, and provide their feedback in terms of likability. Subject responses are divided into three categories: Like, Neutral and Dislike. A novel multimodal system is developed using the developed dataset. The model learns feature representation from data based on the three provided categories. The proposed system contains ensemble of time distributed convolutional neural network, 3D convolutional neural network, and long short term memory networks. All the modalities in proposed architecture are evaluated independently as well as in distinct combinations. The paper also provides detailed insight into learning behavior of the proposed system.
Deep learning methods for the super-resolution problem are showing great performance compared to other traditional techniques. However, these methods are unable to learn complex spatial structures and high frequency details; which leads to over-smooth results. In the present paper, a novel Generative Adversarial Network based architecture named as Residue and Semantic feature based Dual Subpixel Generative Adversarial Network has been proposed for generator and discriminator networks to solve super-resolution problem. The generator network is residue and semantic feature based dual subpixel generative architecture. This architecture is divided into two stages: premier residual stage and deuxieme residual stage. These two stages are concatenated together to form a two stage upsamping process, which enhances the feature learning capability of our model. Inter and intra residual connections are made within these two stages; helping us to sustain the high texture details of images. Semantic based information is implanted in generator to enhance the quality of objects in an image. For embedding semantic information in generator, feature maps extracted from pre-trained model are merged with the input image. To stabilize the training process, we introduced spectral normalization in the discriminator. Visual perception and mean opinion score shows that proposed method outperforms the other state-of-the-art methods.
This paper presents ultra wide band low noise amplifier (LNA) with three stages for 1.9 to 8.2 GHz frequency band based on CMOS technology. In this LNA, first stage is designed on current reuse topology to achieve higher gain, while second stage is complementary push pull configuration in self biased state, which provides high gain and low noise contribution and the third stage is common source stage. This paper proposes two different designs. These designs are optimized for different set of parameters, which makes them suitable for two different categories of applications. The first design with single power source, is able to achieve gain of 21dB with average noise figure of 1.41 dB. Also, 3rd order intercept point (IIP3) and 1-dB compression point are 6 dBm and -17 dBm respectively. While second design achieves reduction in the power dissipation and enhances input output impedance matching. Power dissipation of the second design is 4.1 mW only. Both Ultra Wide Band Low Noise Amplifiers (UWBLNA) are designed and simulated using TMSC 130 nm CMOS technology.
Advanced driver assistance systems (ADAS) have been developed to automate and modify vehicles for safety and better driving experience. Among all computer vision modules in ADAS, 360-degree surround view generation of immediate surroundings of the vehicle is very important, due to application in on-road traffic assistance, parking assistance etc. This paper presents a novel algorithm for fast and computationally efficient transformation of input fisheye images into required top down view. This paper also presents a generalized framework for generating top down view of images captured by cameras with fish-eye lenses mounted on vehicles, irrespective of pitch or tilt angle. The proposed approach comprises of two major steps, viz. correcting the fish-eye lens images to rectilinear images, and generating top-view perspective of the corrected images. The images captured by the fish-eye lens possess barrel distortion, for which a nonlinear and non-iterative method is used. Thereafter, homography is used to obtain top-down view of corrected images. This paper also targets to develop surroundings of the vehicle for wider distortion less field of view and camera perspective independent top down view, with minimum computation cost which is essential due to limited computation power on vehicles.