The results obtained by many current low light image enhancement algorithms exhibit unnatural effects and insufficient contrast. This paper proposes a low light image enhancement algorithm using a fusion-based and multi-layer model. We decompose the illumination of the input image into high-frequency and low-frequency components. Then, we adjust the illumination using two different correcting functions based on a priori statistics and decompose them iteratively until the final distribution is uniform. Finally, we obtain enhanced results by multiplying all the decomposed reflectance with the last layer of corrected illumination. The experiments show that our proposed method achieves excellent image naturalness, contrast, and quality.

Low-light image enhancement plays a significant role in image processing and analysis. Nevertheless, most current algorithms are suffering from the problem of color distortion. With the assistance of an image captured in normal illumination as a reference, the matter of color distortion can be effectively settled by means of the color transfer method. The enhanced images in our proposed algorithm are much closer to the color distribution of the original objects under normal illumination. Our proposed algorithm takes a low-light image and a reference image with a different viewpoint or similar scene as input. Two images are both divided into regions by FCM (Fuzzy C-means). Accordingly, they are then transferred into HSV color space, and the optimal matched results would be obtained by minimizing the weighted distance of the H and S channels. Eventually, the original image would be significantly enhanced using our improved color transfer algorithm. Therefore, this novel low-light image enhancement algorithm based on information assistance expands the application of color transfer in the field of low-light image enhancement. In conclusion, the proposed algorithm has made superior achievements in image contrast retention, color consistency, and naturalness preservation, which would be promising in broad application scenarios.

The naturalness of warps is gaining extensive attention in image stitching. Recent warps, such as SPHP and AANAP, use global similarity warps to mitigate projective distortion (which enlarges regions); however, they necessarily bring in perspective distortion (which generates inconsistencies). In this paper, we propose a novel quasi-homography warp, which effectively balances the perspective distortion against the projective distortion in the non-overlapping region to create a more natural-looking panorama. Our approach formulates the warp as the solution of a bivariate system, where perspective distortion and projective distortion are characterized as slope preservation and scale linearization, respectively. Because our proposed warp only relies on a global homography, it is thus totally parameter free. A comprehensive experiment shows that a quasi-homography warp outperforms some state-of-the-art warps in urban scenes, including homography, AutoStitch and SPHP. A user study demonstrates that it wins most users' favor, compared to homography and SPHP.

The minimum all-ones problem was shown to be NP-complete for general graphs. Therefore, it becomes an interesting problem to identify special classes of graphs for which one can find polynomial time algorithms. In this paper we consider this problem for trees. First, for any solution to the all-ones problem for a tree, we give a characterization of the elements in the solution by introducing the concept of the quasi all-ones problem. Then we give the enumeration for the number of solutions in a tree. By using the minimum odd (even) sum problem as subprocess, we obtain a linear time algorithm for the minimum all-ones problem for trees. We also get a linear time algorithm for finding solutions to the all-ones problem in a unicyclic graph.