High Efficiency Video Coding (HEVC) is the latest video coding standard. The residual quadtree (RQT) coding structure which provides variable block size for transform coding, is adopted in HEVC to achieve high coding efficiency. However, compared with previous standards, encoding complexity is increased significantly in HEVC due to the advanced encoding structure. A fast transform unit (TU) mode decision algorithm using residual difference is proposed in this paper to reduce the computational complexity. The proposed algorithm utilized the residual difference to determine the criterion of early TU termination and early TU skip. The threshold was trained from the beginning samples of each sequence. Experimental results showed that the proposed algorithm saves up to 75.64% and on average 64% TU encoding time compared with HM 15.0 in low delay P configuration and the loss of average BD-BR is less than 0.5%.

The latest High Efficiency Video Coding (HEVC) standard achieves about 50% bit rate saving while maintaining the same subjective quality compared to H.264/AVC High Profile. However, the better coding efficiency is obtained at the cost of significantly increased encoder complexity. In this paper, a fast mode decision algorithm is proposed to relieve the computational burden. The proposed algorithm consists of two steps. Firstly, depth information of CU block is used to reduce the mode candidates, with the assumption that the exhaustive search for a large CU is unnecessary. Secondly, a directional ratio is applied to estimate the rough orientation of each CU, thus some unlikely selected directional modes can be further eliminated. On average, the proposed algorithm can achieve 34.1% encoder time saving while cause an negligible coding performance loss under the All-Intra configuration compared with HM 16.0.

This paper presents an efficient algorithm for motion estimation to reduce High Efficiency Video Coding (HEVC) standard encoding complexity. Phase correlation is initially utilized as a preprocessing step to indicate an approximation of the shift between coding units in the current frame and the reference frame. This is followed by a 9-point diamond search centered on the shift found in the initial step, in order to refine the best matching block. The proposed method has the potential to yield substantial improvements in terms of execution time and resulting video quality in comparison to the traditional search methods.

This paper investigates the compression of infrared images with three codecs: JPEG2000, JPEG-XT and HEVC. Results are evaluated in terms of SNR, Mean Relative Squared Error (MRSE) and the HDR-VDP2 quality metric. JPEG2000 and HEVC perform fairy similar and better than JPEG-XT. JPEG2000 performs best for bits-per-pixel rates below 1.4 bpp, while HEVC obtains best performance in the range 1.4 to 6.5 bpp. The compression performance is also evaluated based on maximum errors. These results also show that HEVC can achieve a precision of 1°C with an average of 1.3 bpp.

Sparse coding - modelling data vectors as sparse linear combinations of basis elements - has been widely and successfully used in image classification, noise reduction, texture synthesis, audio processing, etc. Although traditional sparse coding with fixed dictionaries like wavelet and curvelet can produce promising results, unsupervised sparse coding has shown its advantage by optimizing the dictionary based on target data provided. However, most of the existing unsupervised sparse coding method failed to consider the high dimensional manifold information. Recently, graph regularized sparse coding has been proposed to incorporate manifold information. Better classification and clustering results have been shown compared with naive unsupervised sparse coding. The authors utilize modified feature-sign search and Lagrange dual algorithm to solve the objective function as two consecutive convex functions. This method relies on large number of iterations to get state-of-art classification and clustering results, which is computational intensive. In this paper, we proposed a novel modified online dictionary learning method which iteratively utilizes modified least angle regression and block coordinate descent method to solve the problem. Instead of getting entire coefficient matrix then generate dictionary matrix, our method updates coefficient vector and dictionary matrix in each inner iteration. Thus, efficiency and accuracy are reserved at same time.

In this study, we develop an unsupervised coarse-to-fine video analysis framework and prototype system to extract a salient object in a video sequence. This framework starts from tracking grid-sampled points along temporal frames, typically using KLT tracking method. The tracking points could be divided into several groups due to their inconsistent movements. At the same time, the SLIC algorithm is extended into 3D space to generate supervoxels. Coarse segmentation is achieved by combining the categorized tracking points and supervoxels of the corresponding frame in the video sequence. Finally, a graph-based fine segmentation algorithm is used to extract the moving object in the scene. Experimental results reveal that this method outperforms the previous approaches in terms of accuracy and robustness.

Selecting regions of interest (ROI) of the medical images is an important task in medical image processing. Manual selection of ROIs serves as the main method for single images and it has a high accuracy. However, it will become infeasible to manually segment ROIs on a large number of images. Observing this problem, this paper proposes a fast and accurate segmentation method to obtain ROIs on a batch of medical images. Firstly, we segment the standard brain image S_t which has not been injected with tracer. Secondly, we use a B-Spline elastic registration method to get the inverse-registration parameters. Thirdly, we get the template image T_e with the registration parameters. Finally, we search the target region by template matching. Experimental results show that the proposed method performs well on medical image segmentation.

In this paper, we propose an accurate and robust video segmentation method. The main contributions are threefold: (1) multiple cues (appearance and shape) are explicitly used and adaptively combined to determine segment probability; (2) motion is implicitly used to compute the shape cue; and (3) the segment labeling is improved by utilizing geodesic graph cuts. Experimental results show the effectiveness of the proposed method. © 2016 Society for Imaging Science and Technology.

In this paper, for infrared images, the image enhancement technique based on wavelet transform is studied, which is a process that automatically apply different filtering coefficient toward different directions. The algorithm, including the application of nonlinear anisotropic diffusion, is experienced to the enhancement of infrared images. For directional filtering, the structural feature at each pixel is analyzed by the eigen-analysis. If the analysis shows that the pixel belongs to the edge region, we then perform directional smoothing along the tangential direction of the edge to improve its continuity, while directional sharpening along the normal direction to enhance the contrast. Meanwhile, the noise in the homogeneous region has been reduced notably by applying the appropriate wavelet coefficient. The algorithm is so effective that it reduces the noise while enhancing the edge sharpness at the same time. The quantitative measurements along with the visual inspection were also compared and results showed the algorithm based on wavelet transform has the ability in enhancing the infrared image. The proposed algorithm is compared to the other regular noise-reducing algorithms. The experimental results show that the proposed algorithm considerably improves the infrared image quality without causing any noticeable artifacts. Out of the algorithms compared, our algorithm demonstrated the best performance.