Yung-Yao Chen

3D model retrieval based on deep Autoencoder neural networks

The rapid growth of 3D model resources for 3D printing has created an urgent need for 3D model retrieval systems. Benefiting from the evolution of hardware devices, visualized 3D models can be easily rendered using a tablet computer or handheld mobile device. In this paper, we present a novel 3D model retrieval method involving view-based features and deep learning. Because 2D images are highly distinguishable, constructing a 3D model from multiple 2D views is one of the most common methods of 3D model retrieval. Normalization is typically challenging and time-consuming for view-based retrieval methods; however, this work utilized an unsupervised deep learning technique, called Autoencoder, to refine compact view-based features. Therefore, the proposed method is rotation-invariant, requiring only the normalization of the translation and the scale of the 3D models in the dataset. For robustness, we applied Fourier descriptors and Zernike moments to represent the 2D features. The experimental results testing our method on the online Princeton Shape Benchmark Dataset demonstrate more accurate retrieval performance than other existing methods.

A Spatial-Pyramid Scene Categorization Algorithm based on Locality-aware Sparse Coding

Scene recognition has a wide range of applications, such as object recognition and detection, content-based image indexing and retrieval, and intelligent vehicle and robot navigation. In particular, natural scene images tend to be very complex and are difficult to analyze due to changes of illumination and transformation. In this study, we investigate a novel model to learn and recognize scenes in nature by combining locality constrained sparse coding (LCSP), Spatial Pyramid Pooling, and linear SVM in end-to-end model. First, interesting points for each image in the training set are characterized by a collection of local features, known as codewords, obtained using dense SIFT descriptor. Each codeword is represented as part of a topic. Then, we employ LCSP algorithm to learn the codeword distribution of those local features from the training images. Next, a modified Spatial Pyramid Pooling model is employed to encode the spatial distribution of the local features. For the final stage, a linear SVM is employed to classify local features encoded by Spatial Pyramid Pooling. Experimental evaluations on several benchmarks well demonstrate the effectiveness and robustness of the proposed method compared to several state-of-the-art visual descriptors.

Multi-cue pedestrian detection from 3D point cloud data

Pedestrian detection is one of the key technologies of driver assistance system. In order to prevent potential collisions, pedestrians should be always accurately identified whether during the day or at night. Since the visual images of the night are not clear, this paper proposes a method for recognizing pedestrians by using a high-definition LIDAR without visual images. In order to handle the long-distance sparse point problem, a novel solution is introduced to improve the performance. The proposed method maps the three-dimensional point cloud to the two-dimensional plane by a distance-aware expansion approach and the corresponding 2D contour and its associated 2D features are then extracted. Based on both 2D and 3D cues, the proposed method obtains significant performance boosts over state-of-the-art approaches by 13% in terms of F1-measure.

Design of image barcodes for future mobile advertising

Mobile advertising refers to communication in which mobile phones are used as a medium to efficiently attract potential customers. Among mobile advertising applications, barcodes are becoming a very powerful mobile commerce tool. By capturing a barcode with a camera scanner, people can easily access a wealth of information online. Barcodes have thus converted hard copies of newspapers, wallpapers, and magazines into crucial platforms for mobile commerce. However, although barcodes are frequently used for embedding information in printed matter, they have unsightly overt patterns. Concealing data in visually meaningful image barcodes (such as trademarks) instead of using extra barcode areas has the advantage of increasing the added value of using conventional barcode patterns, and thus it is desirable for future mobile advertising. This paper presents a novel data-hiding method for halftone images. Data-hiding and halftoning algorithms are integrated into the method to prevent extreme bi-level quantization in the halftoning process, for printing purposes. We first applied regular ordered dithering to divide the input continuous-tone image into discrete-parameter halftone cells. The hidden data message was embedded in the halftone cells by using a screen column-shift method. A modified direct binary search (DBS) optimization framework was then used to enhance the output image quality. Simulation results demonstrated that the proposed method outperformed a state-of-the-art method in terms of image quality and data capacity.

Pedestrian characterization in infrared thermography imaging

This paper characterizes the pedestrian by an infrared thermal imaging system. The authors designed and conducted experiments to measure different components which, in general, compose a pedestrian. The experimental results can be used to further interpret thermal images when thermal sensors are applied for pedestrian detection.

A hybrid image enhancement method

For enhancing image contrast, global histogram equalization uses the histogram information of the entire input image for designing its transformation function. However, such global approach is suitable for overall enhancement, and it fails to adapt with the local image brightness features. In this paper, we present an effective method for image contrast enhancement that combines local information and global information. In local enhancement, we get multiple intensity mapping functions from the recursive mean-separate histogram equalization method. According to intensity level, we map different transformation functions to the center sub-block and its neighboring sub-blocks. This method is combined with unsharp masking to enhance the local detail of the image. Experimental results show the effectiveness of the proposed method.

A Robust Learning-Based Detection and Tracking Algorithm

Object tracking in video is a challenging problem in several applications such as video surveillance, video compression, video retrieval, and video editing. Tracking an object in a video is not easy due to loss of information caused by illumination changing in a scene, occlusions with other objects, similar target appearances, and inaccurate tracker responses. In this paper, we present a novel object detection and tracking algorithm via structured output prediction classifier. Given an initial bounding box with its position, we first divide it into sub-blocks with a predefined size. Next, we extract the features from each sub-blocks with Haar-like features method. And then we learn those features with a structured output prediction classifier. We treat the sub-blocks obtained from the initial bounding box as positive samples and then randomly choose negative samples from search windows defined by the specific area around the bounding box. After that, we obtain prediction scores for each sub-blocks both from positive and negative samples. We construct a region-graph with sub-blocks as nodes and classifier’s score as weight to detect the target object in each frame. Our experimental results show that the proposed method outperforms state-of-the-art object tracking algorithms.

Intraframe Coding with Massive Dictionaries of Tree-Structured Representations

H.264/AVC belongs to block-based coding category. For the H.264/AVC intraframe coding, a frame is first divided into non-overlapped blocks, and then intra prediction and the DCT-like transform are applied block by block. For the inter frame coding, H.264/AVC employs variable-block-size for motion compensation and transform stages that can significantly improve the coding performance compared with previous video coding standards. However, for the transform stage of intraframe coding, H.264/AVC only provides two fixed transform block sizes: 4 × 4 and 8 × 8 blocks. In this paper, we propose to employ two massive dictionaries (dyadic and multitree) of representations to improve the efficiency of H.264/AVC intraframe coding. In addition, the corresponding rate-distortion cost function and a fast search algorithm are introduced to incorporate the massive dictionaries into the transform stage. The experimental results show that the proposed method outperforms H.264/AVC in terms of both subjective and objective evaluations.