Ming-Ming Cheng (程明明)

I am a fourth year Ph.D Student in Tsinghua University, under supervision of Prof. Shi-Min Hu. Before that, I received my bachelor degree from Xidian University in 2007. I'm now working in the CSCG lab.

My research interests include: image and video processing, computer vision and machine learning. I am funded by the Google PhD Fellowship, IBM PhD Fellowship and “New PHD Researcher Award” from Chinese Ministry of Education.

Phone: 86-10-62797001-807 (Office)
Address: FIT Building 3-523, Tsinghua University, Beijing, P.R. China, P.C: 100084.

Abstract: Reliable estimation of visual saliency allows appropriate processing of images without prior knowledge of their content, and thus remains an important step in many computer vision tasks including image segmentation, object recognition, and adaptive compression. We propose a regional contrast based saliency extraction algorithm, which simultaneously evaluates global contrast differences and spatial coherence. The proposed algorithm is simple, efficient, and yields full resolution saliency maps. Our algorithm consistently outperformed existing saliency detection methods, yielding higher precision and better recall rates, when evaluated using one of the largest publicly available data sets. We also demonstrate how the extracted saliency map can be used to create high quality segmentation masks for subsequent image processing. [Project page] [C++] [Bib]

Abstract: Connectedness of random walk segmentation is examined, and novel properties are discovered, by considering electrical circuits equivalent to random walks. A theoretical analysis shows that earlier conclusions concerning connectedness of random walk segmentation results are incorrect, and counterexamples are demonstrated.[Matlab Code] [Pdf] [Bib]

Abstract: Repeated elements are ubiquitous and abundant in both manmade and natural scenes. Editing such images while preserving the repetitions and their relations is nontrivial due to overlap, missing parts, deformation between instances, illumination variation, etc. Manually enforcing such relations is laborious and error prone. We propose a novel framework where simple user input in the form of scribbles are used to guide detection and extraction of such repeated elements. Our detection proceebss is based on a novel boundary band method, and robustly extracts the repetitions along with their mutual depth relations. We then use topological sorting to establish a partial depth ordering of overlapping repeated instances. Missing parts on occluded instances are completed using information from other instances. The extracted repeated instances can then be seamlessly edited and manipulated for a variety of high level tasks that are otherwise difficult to perform. We demonstrate the versatility of our framework on a large set of inputs of varying complexity, showing applications to image rearrangement, edit transfer, deformation propagation, and instance replacement. [Project page] [Bib]

Abstract: We present a system that composes a realistic picture from a user provided sketch with text labels. The composed picture is generated by seamlessly stitching several photographs automatically searched from internet according to the sketch and its text labels. While on line image search generates noisy results, our system can automat ically select suitable photographs to generate a high quality com position. To achieve this, we first design a filtering scheme to exclude undesirable images from searched results. Then we propose a novel image blending algorithm for seamless image composition. Our blending algorithm returns a numeric score for each blending, which is used to optimize the combination of searched images. Several vivid results are generated in the experiments. We also perform a user study to demonstrate the advantages of our system. [Project page] [Bib] [PPT] [Commercial Software]

Note: This work is selected as one of the 10 most innovative and promising worldwide initiatives of 2009 by the Netexplorateur jury, under the aegis of French government. It’s also reported by several famous international media, including UK Telegraph and Spiegel. The related introduction video in Vimeo has attracted 1,000,000 clicks. I'm in charge of the image filter part in this work.

Abstract:We present a novel image resizing method which attempts to ensure that important local regions undergo a geometric similarity transformation, and at the same time, to preserve image edge structure. To accomplish this, we define handles to describe both local regions and image edges, and assign a weight for each handle based on an importance map for the source image. Inspired by conformal energy, which is widely used in geometry processing, we construct a novel quadratic distortion energy to measure the shape distortion for each handle. The resizing result is obtained by minimizing the weighted sum of the quadratic distortion energies of all handles. Compared to previous methods, our method allows distortion to be diffused better in all directions, and important image edges are well-preserved. The method is efficient, and offers a closed form solution. [Pdf] [PPT] [Video] [Supplemental] [Bib] [Code]

Abstract:We propose a novel method for curve structure extraction of cartoon images. Our method handles two types of cartoon curves, decorative curves and boundary curves, in a uniform way. The method consists of two steps. First, we calculate curve points by applying non-maximal suppress on secondary derivative of cartoon images. Second, these curve points are linked together to form structure curves while unreliable curves are removed away. Compared to curve structure extraction algorithm proposed by Steger, the number of curves generated by our algorithm is only 19% of Steger’s on average, with better curve quality. Furthermore, more accurate curve position can be obtained by our method. [Doc] [Pdf] [C++ code]

Abstract:Public lighting consumes 30% of the amount of electric power in the entire lighting system and seems to be a growing tendency of bottleneck in energy conservation. In this paper, Starswave, an intelligent street lamp control system, is present to address this problem. It incorporates LEDs to replace the traditional lamps, utilizes the sensors, and controls light switches according to the arrival and departure of vehicles in an intelligent way. In addition, it collects and analyzes the related information from various sources, such as Internet and weather broadcast to improve the management and maintenance of lamps. This paper first describes the design of system components and their implementation, and then provides an algorithm description in details. Of particular concern is the way in which both components and algorithms work together. Further, an extensive experiment study was conducted and the results show that Starswave achieves both high system performance and energy conservation, and outcomes the traditional public lighting facilities significantly. [Pdf] [PPT] [FLASH]

MM Cheng, WB Li, L Qiu, SF Zhang, D Zhou. Starswave: an Intelligent Street Lamp System, in Windows Embedded Student Challenge, 2006. (Third place in world final)