Fast retrieval of images from database is done by unsupervised image categorization technique. CBIR effectiveness is based on the image categorization. For image categorization technique, the image features are extracted by using Scale Invariant Feature Transform (SIFT). Image Categorization and Content-Based Image Retrieval (CBIR) allows automatic extraction of target images according to object feature contents of the image itself. Haar Transform is used to decompose color images into multilevel scale. D4 wavelet Transform is used for the conversion of wavelet coefficients. A progressive image retrieval strategy is achieved by flexible CBIR. In terms of recall rate and retrieval speed, the retrieval performance of D4 and

We present a system and a method to segment the head-shoulder image of participants in a video call using mobile devices such as a smartphone or a tablet. Participants can choose to send only the segmented headshoulder foreground image and overlay it on top of a static background image or a background video on the receiver side of the video call, as well as to replace the background of the caller himself/herself during the video call. Our proposed method extracts the head-shoulder area of each video frame based on detected face region, superpixel clustering, and efficient label propagation.

Visual recognition systems for videos using statistical learning models often show degraded performance when being deployed to a real-world environment, primarily due to the fact that training data can hardly cover sufficient variations in reality. To alleviate this issue, we propose to utilize the object correspondences in successive frames as weak supervision to adapt visual recognition models, which is particularly suitable for human profile recognition. Specifically, we substantialize this new strategy on an advanced convolutional neural network (CNN) based system to estimate human gender, age, and race. We enforce the system to output consistent and stable results on face images from the same trajectories in videos by using incremental stochastic training. Our baseline system already achieves competitive performance on gender and age estimation as compared to the state-of-the-art algorithms on the FG-NET database. Further, on two new video datasets containing about 900 persons, the proposed supervision of correspondences improves the estimation accuracy by a large margin over the baseline.

We present a framework for unsupervised image categorization in which images containing specific objects are taken as vertices in a hypergraph and the task of image clustering is formulated as the problem of hypergraph partition. First, a novel method is proposed to select the region of interest (ROI) of each image, and then hyperedges are constructed based on shape and appearance features extracted from the ROIs. Each vertex (image) and its k-nearest neighbors (based on shape or appearance descriptors) form two kinds of hyperedges. The weight of a hyperedge is computed as the sum of the pairwise affinities within the hyperedge. Through all of the hyperedges, not only the local grouping relationships among the images are described, but also the merits of the shape and appearance characteristics are integrated together to enhance the clustering performance. Finally, a generalized spectral clustering technique is used to solve the hypergraph partition problem. We compare the proposed method to several methods and its effectiveness is demonstrated by extensive experiments on three image databases.

To determine the root cause of through-hole cracking in a flange, detailed methods, including macroscopic and microscopic inspection, microstructural investigations, chemical analysis, hardness measurement, simulation experimentation, and analytic calculations were carried out, in this study. The results show that the cracking was due to overload which was caused by impact. The crack formed in the time period before chromic acid anodization and machining, and the impact stress originated from conveying and revolving transportations. The microstructure, hardness, and chemical compositions of intermediate casing flange all coincide with technical requirements.

Most research efforts on image classification so far have been focused on medium-scale datasets, which are often defined as datasets that can fit into the memory of a desktop (typically 4G~48G). There are two main reasons for the limited effort on large-scale image classification. First, until the emergence of ImageNet dataset, there was almost no publicly available large-scale benchmark data for image classification. This is mostly because class labels are expensive to obtain. Second, large-scale classification is hard because it poses more challenges than its medium-scale counterparts. A key challenge is how to achieve efficiency in both feature extraction and classifier training without compromising performance. This paper is to show how we address this challenge using ImageNet dataset as an example. For feature extraction, we develop a Hadoop scheme that performs feature extraction in parallel using hundreds of mappers. This allows us to extract fairly sophisticated features (with dimensions being hundreds of thousands) on 1.2 million images within one day. For SVM training, we develop a parallel averaging stochastic gradient descent (ASGD) algorithm for training one-against-all 1000-class SVM classifiers. The ASGD algorithm is capable of dealing with terabytes of training data and converges very fast-typically 5 epochs are sufficient. As a result, we achieve state-of-the-art performance on the ImageNet 1000-class classification, i.e., 52.9% in classification accuracy and 71.8% in top 5 hit rate.

The traditional SPM approach based on bag-of-features (BoF) requires nonlinear classifiers to achieve good image classification performance. This paper presents a simple but effective coding scheme called Locality-constrained Linear Coding (LLC) in place of the VQ coding in traditional SPM. LLC utilizes the locality constraints to project each descriptor into its local-coordinate system, and the projected coordinates are integrated by max pooling to generate the final representation. With linear classifier, the proposed approach performs remarkably better than the traditional nonlinear SPM, achieving state-of-the-art performance on several benchmarks. Compared with the sparse coding strategy [22], the objective function used by LLC has an analytical solution. In addition, the paper proposes a fast approximated LLC method by first performing a K-nearest-neighbor search and then solving a constrained least square fitting problem, bearing computational complexity of O(M + K2). Hence even with very large codebooks, our system can still process multiple frames per second. This efficiency significantly adds to the practical values of LLC for real applications.

The paper presents a failure case of an aero-engine combustor liner. Macroscopic and microscopic observations, microstructure investigation, chemical analysis and hardness measurement have been performed to investigate the damage mechanism and failure causes of notches and cracks. The results show that the failure modes of notches, axial-direction crack along the combustor liner and crack at the edge of welding spot are respectively attributed to high-temperature ablation, mixed crack of ablation and thermal fatigue crack. The bad configuration of crossover tube is the main reason for the failure of combustor liner.

This notebook paper summarizes Team NEC-UIUC’s approaches for TRECVid 2010 Evaluation of Semantic Indexing. Our submissions mainly take advantage of advanced image classification methods using linear coordinate coding (LCC) of local features powered by the distributed computing software Hadoop. For every video shot, we evenly sample key frames and extract dense local features including DHOG and LBP, which are encoded by linear coordinate coding. Then, for every concept large-scale linear SVM classifiers are trained based on spatial pyramid of LCC features. Finally, we employ multiple instance learning to rank the video shots according to the SVM scores of individual frames. Our systems achieve mean extended inferred average precision (mean xinfAP) 7.40% for the 30 concepts evaluated by NIST and mean average precision 28.63% using 1/5 of the development data as the validation set for the total 130 concepts.