In this paper a new method for determining the search area for Motion Estimation algorithm based on block matching is suggested. In the proposed method the search area is adaptively found for each block of a frame. This search area is similar to that of the full search (FS) algorithm but smaller for most blocks of a frame. Therefore, the proposed algorithm is analogous to FS in terms of regularity but has much less computational complexity. To find the search area, the temporal and spatial correlations among the Motion vectors of blocks are used. Based on this, the matched block is chosen from a rectangular area that the prediction vectors set out. Simulation results indicate that the speed of the proposed algorithm is at least 7 times better than the FS algorithm.      

An important technique in image and video processing is global Motion Estimation (GME). The common GME methods can be classified in direct and indirect categories. Whereas the direct global Motion Estimation techniques boast reasonable precision they tend to suffer from high complexity. As with indirect methods, though presenting lower complexity, they mostly exhibit lower accuracy than their direct counterparts. In this paper, the authors introduce a robust algorithm for GME with near identical accuracy and almost 50-times faster than MPEG-4 verification model (VM). This approach entails two stages in which, first, Motion vector of sampled block is employed to obtain initial GME then Levenberg-Marquardt algorithm is applied to the subsampled pixels to optimize the initial GME values. As will be shown, the proposed solution exhibits remarkable accuracy and speed features with experimental results distinctively bearing them out.

Motion Estimation and compensation is an essential part of existing video coding systems. The mesh-based Motion Estimation (MME) produces smoother Motion field, better subjective quality (free from blocking artifacts), and higher peak signal-to-noise ratio (PSNR) in many cases, especially at low bitrate video communications, compared to the conventional block matching algorithm (BMA). However, the iterative refinement process of MME is computationally much costly and makes the method impractical in real- (or near real-) time systems. Also, eliminating the iterative refinement step deteriorates the Motion Estimation result. In this paper, we propose Motion adaptive interpolation schemes for non-iterative MME, which use BMA to compute the Motion vectors (MVs) of mesh nodes. The proposed algorithm aims at compromising the MME and BMA by modifying the interpolation patterns (IPPs) of the MME in an adaptive manner, based on the MVs of mesh nodes. Experimental results show notable rate-distortion improvement over both BMA and conventional non-iterative MME, with acceptable visual quality and system complexity, especially when applied to sequences with medium to high Motion activities.    

Background: Myoelectric control is a method of decoding the motor intent from the electromyogram (EMG) data and using the estimated intent to control prostheses and robots. This work investigates Estimation of the wrist kinematics from EMG signals using polynomial models. Due to their low complexity, polynomial models are potentially the perfect choice for EMG-kinematics modeling. Methods: Ten ablebodied individuals participated in this study, where the EMG signals from the forearm and the wrist kinematics from the contralateral wrist were measured during mirrored contractions. Two sets of EMG features were employed including the time domain (TD) set, and TD features along with autoregressive coefficients (TDAR). Polynomial models of order 1 to 4 were applied to map the EMG signals to the wrist Motions. The performance was directly compared to that of a multilayer perceptron (MLP) neural network. Results: The Estimation accuracy of the wrist kinematics improved with increasing the order of the model, but saturated at the 4 th order. When using the TD set, the MLP significantly outperformed all polynomial models. However, when using the TDAR set, the polynomial models’ performance improved so that the 4 th order model performance was not significantly different than that of the MLP in two DoFs, although it was lower than MLP in one DoF. Conclusion: These results indicate that polynomial models are not as effective as more complex models such as neural networks, in learning the highly nonlinear mapping between the EMG data and Motion intent. However, using a sufficiently high number of various EMG features, would reduce the mapping nonlinearities, and thereby may increase the polynomial models’ performance to levels similar to those of complex black box models.

Nowadays, the use of hardware/software codesign has grown dramatically in the design of embedded systems since it can improve the processing power, system efficiency, and the total cost of production. In this method, some parts of the system are implemented in hardware and the other parts are implemented in software in order to satisfy the system constraints, including power consumption, area and processing time. This paper proposes a parallel architecture for Motion Estimation in HEVC encoder. In the proposed method, the Motion Estimation part of the encoder, which has a high computational complexity, is implemented in hardware, and the computational complexity of this part is improved using parallel processing. The hardware implementation of Motion Estimation part is much less complex than the adopted HM reference software, making it more suitable for embedded systems. Experimental results show a significant improvement over software implementation.