Abstract:Object detection and recognition is a hot topic in computer vision.Traditional methods use only local features for detection.Recently,some research results show that the detecting performance could be improving by using topic features.Some researchers employed topic models which is originally used for text analysis to extract topic features from images for object recognition and detection.However,visual features should be quantized into virtual words and information of class label should be ignored while using traditional topic models such as probabilistic latent sematic analysis(PLSA),latent Dirichlet allocation(LDA) and so on.In order to utilize continuous local features and information of class label in one model,we propose a new graphical model named latent Dirichlet classification(LDC),which is inspired by LDA model.The proposed model has three more variables than LDA in the graphical structure: x(local features),c(class label) and v(parameter).In the proposed model,we consider class label of each image block is determined by both of its local features and topic features based on original LDA model.Parameter v is a set of classifiers trained for combining these two features.Similar with the inference process of LDA model,we use variational inference to solve our model.As a result of continuous local features,information of class label and topic features are all token into consideration reasonably,LDC can be used in object detection directly and efficiently.In the end of this paper,we test the availability of LDC on two datasets.Experimental results show that our proposed model improve the performance of object detection efficiently.

Scene Classification Using Class-Supervised Local-Space-constraint Latent Dirichlet Allocation

Local–class–shared–topic Latent Dirichlet Allocation Based Scene Classification

Hybrid Generative/discriminative Scene Classification Strategy Based on Latent Dirichlet Allocation for High Spatial Resolution Remote Sensing Imagery.

Category Constrained Learning Model for Scene Classification

Incorporate Spatial Information into pLSA for Scene Classification

Latent Dirichlet classification:A new method for object detection

Spatial-DiscLDA for visual recognition

Supervised LDA for Image Annotation

Discovering Class-Specific Spatial Layouts for Scene Recognition

Semantic image classification using statistical local spatial relations model

Scene and place recognition using a hierarchical latent topic model.

Unsupervised Object Localization With Latent Dirichlet Allocation

Supervised labeled latent Dirichlet allocation for document categorization

Supervised Latent Dirichlet Allocation with a Mixture of Sparse Softmax

Scene classification using a hybrid generative/discriminative approach

LDA Model Combined Spatial Information for Visual Object Recognition Research

Bag of Spatial Visual Words Model for Scene Classification

Beyond Local Image Features: Scene Calssification Using Supervised Semantic Representation

Projection Learning with Local and Global Consistency Constraints for Scene Classification

Local Manifold Sparse Model for Image Classification.

Pose Estimation Based on Semi-Supervised Latent Dirichlet Allocation