Incremental kernel learning for active image retrieval without global dictionaries

Abstract

In content-based image retrieval context, a classic strategy consists in computing off-line a dictionary of visual features. This visual dictionary is then used to provide a new representation of the data which should ease any task of classification or retrieval. This strategy, based on past research works in text retrieval, is suitable for the context of batch learning, when a large training set can be built either by using a strong prior knowledge of data semantics (like for textual data) or with an expensive off-line pre-computation. Such an approach has major drawbacks in the context of interactive retrieval, where the user iteratively builds the training data set in a semi-supervised approach by providing positive and negative annotations to the system in the relevance feedback loop. The training set is thus built for each retrieval session without any prior knowledge about the concepts of interest for this session. We propose a completely different approach to build the dictionary on-line from features extracted in relevant images. We design the corresponding kernel function, which is learnt during the retrieval session. For each new label, the kernel function is updated with a complexity linear with respect to the size of the database. We propose an efficient active learning strategy for the weakly supervised retrieval method developed in this paper. Moreover this framework allows the combination of features of different types. Experiments are carried out on standard databases, and show that a small dictionary can be dynamically extracted from the features with better performances than a global one.

A Bayesian beta kernel model for binary classification and online learning problems

“…Because online learning typically does not begin with a large training set that can be used to generate optimal parameters for kernel functions, algorithms have been developed that linearly combine multiple kernels for a single classifier. As the classifier learns on data, the weights for each possible kernel are also updated to improve accuracy . Future research could apply these concepts to the beta kernel model in an online‐learning environment.…”

Re-identificación de personas a través de sus características soft-biométricas en un entorno multi-cámara de video-vigilancia

“…Es por esta razón que se considera el método de kernel alignment (KA para futuras referencias) en este trabajo como el cuarto método para medir la relevancia de cada una de las características extraídas. Con el método del kernel alignment se calculó el alineamiento entre cada característica y el kernel ideal (Gosselin et al, 2011). Cuanto mayor es este alineamiento, mayor se ajustará el kernel a la clase representada por los datos.…”

Soft-biometrics evaluation for people re-identification in uncontrolled multi-camera environments

“…Kernel-based methods are increasingly being used for data modeling because of their conceptual simplicity and outstanding performance in many tasks [27]. Specifically, a kernel target alignment approach has been considered as the fourth ranking method (it will be called KA for future references) [28]. With the kernel alignment method, the alignment between each feature and the ideal kernel is calculated.…”