A Computational Meta-Learning Inspired Model for Sketch-based Video Retrieval

Video retrieval framework based on color co-occurrence feature of adaptive low rank extracted keyframes and graph pattern matching

Video retrieval based on object discovery

This paper proposes a formal framework for image and video retrieval using discrete Markov random fields (MRF). The training dataset consists of images with keywords (regions are not labeled). The model is built using a discrete vocabulary of vector quantized region or point features generated from the training images. Since performance is dependent on the size of the vocabulary, a large vocabulary of a couple of million visterms is used. Such large vocabularies cannot be generated by conventional clustering algorithms so hierarchical k-means is used to generate it. Unlike many previous techniques, our MRF based model doesn't require an explicit annotation step for retrieval. The model directly ranks all test images according to the posterior probability of an image given a query. Traditionally, most models are trained by maximizing likelihood - instead this model is trained by maximizing average precision. Image and video retrieval experiments are performed on two standard datasets (a Corel dataset and a TRECVID3 dataset) which consist of 4,500 images and about 44,100 keyframes respectively. The results show that based on a large visual vocabulary the model runs extremely fast on even very large datasets while having comparable retrieval performance to the best performing (continuous feature) models.

This study examined how performance appraisal systems affect service delivery in the Zambia Police Service, focusing on motivation, accountability, and overall performance. Surveying 100 officers across ranks and stations, the study found that most were aware of appraisal systems, but opinions on fairness and effectiveness varied. Appraisals helped evaluate performance and guide promotions and training, but challenges like weak supervision, politicized culture, limited resources, and rural staffing issues reduced their effectiveness. Innovative strategies such as community feedback, mentoring, and capacity-building improved outcomes. The study concludes that appraisals can enhance service delivery if supported by strong supervision, transparency, minimal political interference, and better digital infrastructure.

Localizing phrases in images is an important part of image understanding and can be useful in many applications that require mappings between textual and visual information. Existing work attempts to learn these mappings from examples of phrase-image region correspondences (strong supervision) or from phrase-image pairs (weak supervision). We postulate that such paired annotations are unnecessary, and propose the first method for the phrase localization problem where neither training procedure nor paired, task-specific data is required. Our method is simple but effective: we use off-the-shelf approaches to detect objects, scenes and colours in images, and explore different approaches to measure semantic similarity between the categories of detected visual elements and words in phrases. Experiments on two well-known phrase localization datasets show that this approach surpasses all weakly supervised methods by a large margin and performs very competitively to strongly supervised methods, and can thus be considered a strong baseline to the task. The non-paired nature of our method makes it applicable to any domain and where no paired phrase localization annotation is available.

Existing Temporal Activity Localization (TAL) methods largely adopt strong supervision for model training, which requires (1) vast amounts of untrimmed videos per each activity category and (2) accurate segment-level boundary annotations (start time and end time) for every instance. This poses a critical restriction to the current methods in practical scenarios where not only segment-level annotations are expensive to obtain, but many activity categories are also rare and unobserved during training. Therefore, Can we learn a TAL model under weak supervision that can localize unseen activity classes? To address this scenario, we define a novel example-based TAL problem called Minimum Effort Temporal Activity Localization (METAL): Given only a few examples, the goal is to find the occurrences of semantically-related segments in an untrimmed video sequence while model training is only supervised by the video-level annotation. Towards this objective, we propose a novel Similarity Pyramid Network (SPN) that adopts the few-shot learning technique of Relation Network and directly encodes hierarchical multi-scale correlations, which we learn by optimizing two complimentary loss functions in an end-to-end manner. We evaluate the SPN on the THUMOS'14 and ActivityNet datasets, of which we rearrange the videos to fit the METAL setup. Results show that our SPN achieves performance superior or competitive to state-of-the-art approaches with stronger supervision.

Remote sensing using unmanned aerial vehicles (UAV) combined with machine learning (ML) has significantly advanced field-scale prediction of aboveground biomass. Although ensemble learning frameworks (ELFs) typically outperform individual ML algorithms in accuracy, systematic evaluations of meta learner selection and the effects of base learner quantity and diversity remain limited. Leveraging vegetation indices and texture features extracted from multi-temporal UAV imagery, ELFs were constructed incorporating nine ML algorithms with three meta learners, Linear model, Random forest and Bayesian model averaging (BMA), to systematically evaluate how base learner configuration affects prediction accuracy. Using the fused feature set of sensitive vegetation indices and texture features, Gaussian process regression (GPR) achieved the highest accuracy among all base learners, with R2 = 0.769 and RMSE = 1.83 t·ha–1. Also, the three meta learners outperformed the best base learner, with the BMA meta learner yielding the superior accuracy (R2 = 0.795, RMSE = 1.73 t·ha–1). However, meta learner performance depended strongly on the composition of the base learners pool, stability was optimal with five base learners, and maximum accuracy was achieved by hybrid ensembles that combined linear-kernel models with GPR. This study highlights the importance of both meta learner selection and base learner composition in ELFs for aboveground biomass prediction in rice. These findings offer methodological guidance for UAV-based high-precision monitoring of crop AGB, with practical implications for precision agriculture and crop management.

Content-based video retrieval is a proper solution to handle the video data. But because of their huge volumes and high dimensionality, finding a proper way to organize them for efficient search and retrieval becomes a challenging and important problem. In this paper, we present a framework for content-based video retrieval using the Shot Cluster Tree, while the latter organizes the content of shots in the tree structure. The framework can supply users with two types of queries: query-by-content and query-by-example. We also put up a novel and efficient clustering method for generating the Shot Cluster Tree structure, where shots which are visual similar and time adjacent are grouped into shot groups using the method of Sliding Shot Window firstly and then shot groups are clustered into shot clusters with the simple agglomerative hierarchical clustering method. In addition to constructing the structure of content, the Shot Cluster Tree also provides better ways to video summary and video annotation, which facilitate the video retrieval. An experimental system has been built up. Experiments verify the effectiveness of the proposed approach.

Affective content recognition is an unsolved technical problem in intricate video retrieval and high-level applications. In this paper , a few video retrieval problems are studied at the view of affective intensity recognition. A novel algorithm about video scene affective intensity recognition, which is based on the unascertained measure, is proposed. The algorithm consists of two parts. Firstly, the sub-module of video affective intensity features extraction introduces the affective intensity feature extraction sub-models of the shots correlation coefficient（SCC） and the duration ratio of the shots（DROS）. Secondly, the sub-module of affective intensity determination includes the unascertained intensity measure matrix of video scene and video affective intensity type recognition sub-model. The experimental results show that the proposed algorithm is feasible and achieves more than 70 percent accuracy.

Motion feature plays an important role in video retrieval. The current literature mostly addresses motion retrieval only by camera motion and global motion of individual video objects in a video scene. In this paper, we propose two new motion descriptors that capture the local motion of the video object within its bounding box. The proposed descriptors are rotation and scale invariant and based on the angular and circular area variances of the video object and the variances of the angular radial transform coefficients. Experiments show that ranking obtained by querying with our proposed descriptors closely match with the human ranking.

The video retrieval is an important component of the personalized information service system. In this paper, a few video retrieval problems are studied at the view of affective content recognition. Based on the unascertained theory, the relation between low-level video emotion features and highlevel cognitive emotion is established, and a novel algorithm about video scene emotion type recognition, which is based on the unascertained measure, is proposed. The algorithm consists of two parts. Firstly, the sub-module of low-level video emotion feature extraction introduces the emotion feature extraction sub-models of the video scene brightness, shot cut rate and color efficacy. Secondly, the sub-module of video emotion determination includes the unascertained measure emotion matrix of video scene and video emotion type recognition submodel. The experimental results show that the proposed algorithm is feasible and achieves more than 80 percent accuracy.

Infrared and visible image fusion based on visual saliency map and weighted least square optimization

The huge collections of unconstrained videos have amplified the so-called semantic gap for content-based video retrieval. Therefore, new efficient approaches with higher generalisation power are needed. In this work, we present an interactive video retrieval approach based on latent topics to cope with the semantic gap in an efficient way. A supervised Symmetric extension of probabilistic Latent Semantic Analysis model is presented (sSpLSA). Then, this model is adapted to an on-line interactive information retrieval problem and it is applied to a video retrieval framework based on explicit short-term Relevance Feedback (RF) where queries are inside the database. Finally, several retrieval simulations using the Consumer Columbia Video (CCV) database are performed to compare the proposed approach with a distance-based RF baseline.KeywordsContent-based video retrievalrelevance feedbacklatent topics

Representing video events is an essential step for a wide range of visual applications. In this paper, we propose the event sketch, a high-level event representation, to depict the dynamic properties of video events composed of actions of semantic objects. We show that this representation can facilitate a novel sketch based video retrieval (SBVR) system, which has not been considered before to the best of our knowledge. In this system, users are allowed to draw the evolutions (e.g. spatiotemporal layouts and behaviors of semantic objects) on a board, and retrieve the events whose semantic objects have the similar evolutions from a database. To do this, event sketches are constructed on both the user queries and database videos, and compared under a novel graph-matching scheme based on data-driven Monta Carlo Markov chain (DDMCMC). To test our approach, we collect a novel dataset of goal events in real soccer videos, which consists actions of multiple players and shows large variability in the evolution process of the events. Experiments on this dataset and the publicly available dataset CAVIAR demonstrated the effectiveness of the proposed approach.

The field of image processing is constantly evolving in the last decades and it is being used in many applications. At present there are many enhancement techniques are available for contrast enhancement of the image. This paper proposes an effective approach to improve the quality of the image based on the image fusion using the discrete wavelet transform (DWT). Contrast enhancement approach is proposed for images which has low contrast. Contrast enhancement involves transforming the intensity of pixels from the original value to new value. The goal of contrast enhancement is to improve visibility of image details without introducing unwanted visual appearances and/or unwanted artefacts. This paper presents the contract enhancement technique to visualize all available information in an image. Proposed method uses the retinex algorithm and gamma correction technique. It uses the DWT for decomposition of the image and decompose the image into a set of band-limited components, called HH, HL, LH, and LL. LL subband will be fused to get the better result of the image. Finally, inverse DWT will be performed on fused image to achieve the final enhanced image.