Age-related macular degeneration (AMD) is a leading cause of vision loss. The presence of drusen are often associated to AMD. Drusen are tiny yellowish-white extracellular buildup present around the macular region of the retina. Clinically, ophthalmologists examine the area around the macula to determine the presence and severity of drusen. However, manual identification and recognition of drusen is subjective, time consuming and expensive. To reduce manual workload and facilitate large-scale early AMD screening, it is essential to detect drusen automatically. In this paper, we propose to use biologically inspired features (BIF) for the purpose of AMD detection. The optic disc and macula are detected to determine a focal region around macula for feature extraction. The extracted features are then classified using support vector machines (SVM). Our experimental results, tested on 350 images, demonstrate that the biologically inspired features from the focal region is effective for drusen detection with a sensitivity of 86.3% and specificity of 91.9%. The results of our proposed approach can be used to reduce workload of ophthalmologists and diagnosis cost.

This paper presents a method of max-pooling spatially-coherent pyramid matching (MpScPM). Higher-layer representations are generated from lower-layer subregions, by a biologically-inspired max pooling strategy. Second, instead of reshaping the pyramid representation into a vector (used in generic SPM), the layer and location information of each subregion are kept and weak geometrical correspondences between matched subregions are explored to enhance our pyramid matching method. To enhance the possibility of finding the best matches at different scales and locations, cross-layer region similarities are computed, while the correspondences (either spatial neighbors or adjacent layers) are also incorporated. We evaluate our proposed MpScPM method on several existing benchmark datasets and it achieves excellent performances.

As there is an exponential increase of web videos, it is time-consuming to get a query result from the tremendous data. An effective and efficient video management system is in urgent need. To increase the efficiency of video retrieval and storage, the most widely used methods are indexing schemes, such as locality sensitive hashing (LSH). However, it is more essential to represent the video itself compactly. In this paper, we propose a strategy to generate stratification-based keyframe cliques (SKCs) for video description, which are more compact and informative than frames or keyframes. The new representations are scalable for different retrieval tasks due to the ranking of SKCs. To further accelerate the retrieval speed, only top SKCs will be used; meanwhile, the searching results are still satisfactory. Experiments are conducted on TRECVID dataset as well as web video dataset. Results show that our proposed SKCs are more succinct and informative for video retrieval and management.

Bag-of-Words is widely used to describe images for image classification. However, this approach is limited because the spatial relation over visual words is not well exploited and also it is difficult to generate a single comprehensive vocabulary. In this paper, we propose novel effective schemes to handle these two issues. First, we propose a structure propagation technique to build more reasonable co-occurrence matrices of visual words to exploit the spatial information, which assigns a higher weight to the co-occurrence over two patches that lie in the same object part. Second, we build the multiple-histogram representation over hierarchical vocabularies to avoid the ambiguity of single vocabulary, and particularly present a learning approach to combine the multiple histograms to integrate both within-vocabulary and cross-vocabulary information. We evaluate our proposed method using the Princeton sports event dataset. Compared to the state-of-the-art results, our proposed approach has shown promising results.

Many multimedia applications can benefit from recognizing image content. It requires a robust and discriminative representation of objects, especially in the situation of only a few training samples available. In this paper, we present a new approach to integrate the advantages of bag-of-words model and part-based model for image recognition. Each image is encoded as a hierarchical word image (HWI), which contains not only visual appearance but also spatial information. The object parts are then located and represented in HWI. Finally, the part-based star model (SM) is used to learn the object model and recognize the test images. It is shown that our proposed approach can detect more accurate part candidates and significantly improve the performance of original part-based model for object recognition.

Current large scale image retrieval engines rely heavily on the surrounding text information, which inevitably includes some irrelevant images in the retrieval results due to the noisy environment. To improve the retrieval performance, we propose an unsupervised web image re-ranking method by incorporating images' visual information. Our method can automatically select a set of representative images from the original image pool as concept model, which is highly related to the query concept and critically important for the re-ranking result. With a similarity graph constructed by top results given by text based retrieval, we utilize Normalized Cut to select the part with the highest similarity density as concept model. We re-rank the rest images according to their similarities to the concept model. The advantages of our method are (i): Our method is unsupervised, and it doesn't need any pre-prepared query/training image or user's feedback, Thus it greatly facilitates users' retrieval. (ii): By finding a set of images rather than single image, we are able to give a more complete and more robust model for the query concept. (iii): Multi-ranking Integration Strategy is adopted to re-rank the rest images. Experiments show that our method can achieve satisfying results.

This paper presents an efficient and effective solution for retrieving image near-duplicate (IND) from image database. We introduce the coherent phrase model which incorporates the coherency of local regions to reduce the quantization error of the bag-of-words (BoW) model. In this model, local regions are characterized byvisual phraseof multiple descriptors instead of visual word of single descriptor. We propose two types of visual phrase to encode the coherency in feature and spatial domain, respectively. The proposed model reduces the number of false matches by using this coherency and generates sparse representations of images. Compared to other method, the local coherencies among multiple descriptors of every region improve the performance and preserve the efficiency for IND retrieval. The proposed method is evaluated on several benchmark datasets for IND retrieval. Compared to the state-of-the-art methods, our proposed model has been shown to significantly improve the accuracy of IND retrieval while maintaining the efficiency of the standard bag-of-words model. The proposed method can be integrated with other extensions of BoW.

This paper presents an efficient and effective solution for retrieving Image Near-Duplicate (IND). Different from traditional methods, we analyze the local dependencies among region descriptors in a spatial-scale space. Such local dependencies in spatial-scale space(LDSS) encodes not only visual appearance but also the spatial and scale co-occurrence of them. The local dependencies are integrated over all spatial locations and multiple scales to form the image representation, which is invariant to spatial transformation and scale change. We evaluate our proposed LDSS method for IND retrieval using an existing benchmark as well as a new dataset extracted from the keyframes of TRECVID corpus. Compared to the state-of-the-art results, local dependencies in spatial-scale space(LDSS) approach has been shown to significantly improve the accuracy of IND retrieval.