Generic Object Classification Research Articles

Rethinking Vision Transformer and Masked Autoencoder in Multimodal Face Anti-Spoofing

Recently, vision transformer (ViT) based multimodal learning methods have been proposed to improve the robustness of face anti-spoofing (FAS) systems. However, there are still no works to explore the fundamental natures (e.g., modality-aware inputs, suitable multimodal pre-training, and efficient finetuning) in vanilla ViT for multimodal FAS. In this paper, we investigate three key factors (i.e., inputs, pre-training, and finetuning) in ViT for multimodal FAS with RGB, Infrared (IR), and Depth. First, in terms of the ViT inputs, we find that leveraging local feature descriptors (such as histograms of oriented gradients) benefits the ViT on IR modality but not RGB or Depth modalities. Second, in consideration of the task (FAS vs. generic object classification) and modality (multimodal vs. unimodal) gaps, ImageNet pre-trained models might be sub-optimal for the multimodal FAS task. Finally, in observation of the inefficiency on direct finetuning the whole or partial ViT, we design an adaptive multimodal adapter (AMA), which can efficiently aggregate local multimodal features while freezing majority of ViT parameters. To bridge these gaps, we propose the modality-asymmetric masked autoencoder (M2\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$^{2}$$\\end{document}A2\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$^{2}$$\\end{document}E) for multimodal FAS self-supervised pre-training without costly annotated labels. Compared with the previous modality-symmetric autoencoder, the proposed M2\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$^{2}$$\\end{document}A2\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$^{2}$$\\end{document}E is able to learn more intrinsic task-aware representation and compatible with modality-agnostic (e.g., unimodal, bimodal, and trimodal) downstream settings. Extensive experiments with both unimodal (RGB, Depth, IR) and multimodal (RGB+Depth, RGB+IR, Depth+IR, RGB+Depth+IR) settings conducted on multimodal FAS benchmarks demonstrate the superior performance of the proposed methods. One highlight is that the proposed method is robust under various missing-modality cases where previous multimodal FAS models suffer serious performance drops. We hope these findings and solutions can facilitate the future research for ViT-based multimodal FAS.

On the robustness of self-supervised representations for multi-view object classification

• Analyse self-supervised and supervised embeddings regarding viewpoint robustness in both controlled and real-world settings. • Controlled setting created by approximating view using a homography on multiple common benchmark datasets. • Performance in-the-wild quantified by analysing the representations on multiple real-world, multi-view datasets. • Demonstrate that self-supervised representations are more robust to both controlled and real-world settings. • Show that self-supervised representations require less context than supervised ones in recognising objects in novel views. It is known that representations from self-supervised pre-training can perform on par, and often better, on various downstream tasks than representations from fully-supervised pre-training. This has been shown in a host of settings such as generic object classification and detection, semantic segmentation, and image retrieval . However, some issues have recently come to the fore that demonstrate some of the failure modes of self-supervised representations, such as performance on non-ImageNet-like data, or complex scenes. In this paper, we show that self-supervised representations based on the instance discrimination objective lead to better representations of objects that are more robust to changes in the viewpoint and perspective of the object. We perform experiments of modern self-supervised methods against multiple supervised baselines to demonstrate this, including approximating object viewpoint variation through homographies , and real-world tests based on several multi-view datasets. We find that self-supervised representations are more robust to object viewpoint and appear to encode more pertinent information about objects that facilitate the recognition of objects from novel views.

A comparitive analysis of mobilenet and xception architecture for classification of endoscopy images

The deep architectures have been making their significant mark in solving the image classification problems. The usage of the standard deep learning architectures has been increasing at a greater pace because of the outcome of the state-of-the art accuracy and precision by incorporating these architectures in image classification problems. The classification of medical images from variety of modalities is also possible with better accuracy by using these deep architectures. The standard architectures that have achieved substantial results in general object classification can be used and their learning can be transferred to the required domain. Further, these architectures can be used in an ensemble structure thereby creating a customized architecture for solving the classification problem. Some of the challenges in including these architectures include, the enormous memory and processing power requirement. Current trends are focused towards generating optimized deep structures that has relatively lesser memory and processing capacity requirement. To this end, the two benchmark architectures that have obtained momentous results in object classification namely the MobileNetV2 and Xception are included in this study to perform binary classification of capsule endoscopy images to detect the presence of abnormality. Capsule endoscopy (CE) is a medical procedure where the patient swallows the pill-based camera that traverses the entire gastrointestinal (GI) path capturing numerous images. These images are to be assessed for detecting the presence of any abnormality. The architectures included in this work have been trained on the images from the publicly available CE datasets and an ablation study on the hyperparameter tuning of the architecture have been conducted. The results are compared, analyzed and presented here. The MobileNetV2 and Xception architecture have achieved a maximum accuracy of 85% and 82% in the abnormality detection in CE images.

PocketNet: Extreme Lightweight Face Recognition Network Using Neural Architecture Search and Multistep Knowledge Distillation

Deep neural networks have rapidly become the mainstream method for face recognition (FR). However, this limits the deployment of such models that contain an extremely large number of parameters to embedded and low-end devices. In this work, we present an extremely lightweight and accurate FR solution, namely PocketNet. We utilize neural architecture search to develop a new family of lightweight face-specific architectures. We additionally propose a novel training paradigm based on knowledge distillation (KD), the multi-step KD, where the knowledge is distilled from the teacher model to the student model at different stages of the training maturity. We conduct a detailed ablation study proving both, the sanity of using NAS for the specific task of FR rather than general object classification, and the benefits of our proposed multi-step KD. We present an extensive experimental evaluation and comparisons with the state-of-the-art (SOTA) compact FR models on nine different benchmarks including large-scale evaluation benchmarks such as IJB-C and MegaFace. PocketNets have consistently advanced the SOTA FR performance on nine mainstream benchmarks when considering the same level of model compactness. With 0.92M parameters, our smallest network PocketNetS-128 achieved very competitive results to recent SOTA compacted models that contain up to 4M parameters. Training codes and pre-trained models are public. <xref ref-type="fn" rid="fn1" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">¹</xref>

Variational Multi-Prototype Encoder for Object Recognition Using Multiple Prototype Images

In the recent research of Variational Prototyping-Encoder (VPE), the problem of classifying 2D flat objects of the unseen class has been addressed. VPE solves this problem by pre-learning the image translation task from real-world object images to their corresponding prototype images as a meta-task. VPE uses a single prototype for each object class. However, in general, a single prototype is not sufficient to represent a generic object class because the appearance can change significantly according to viewpoints and other factors. In this case, using VPE and a single prototype for each class in training can result in overfitting or performance degradation. One solution may be the use of multiple prototypes. However, this also requires costly sub-labeling for dividing the input class into smaller classes and assigning a prototype to each. Therefore, we propose a new learning method, the variational multi-prototype encoder (VaMPE), which can overcome the limitations of VPE and use multiple prototypes for each object class. The proposed method does not require additional sub-labeling other than simply adding multiple prototypes to each class. Through various experiments, we demonstrate that the proposed method outperforms VPE.

Image-Based Scene Recognition for Robot Navigation Considering Traversable Plants and Its Manual Annotation-Free Training

This paper describes a method of estimating the traversability of plant parts covering a path and navigating through them for mobile robots operating in plant-rich environments. Conventional mobile robots rely on scene recognition methods that consider only the geometric information of the environment. Those methods, therefore, cannot recognize paths as traversable when they are covered by flexible plants. In this paper, we present a novel framework of image-based scene recognition to realize navigation in such plant-rich environments. Our recognition model exploits a semantic segmentation branch for general object classification and a traversability estimation branch for estimating pixel-wise traversability. The semantic segmentation branch is trained using an unsupervised domain adaptation method and the traversability estimation branch is trained with label images generated from the robot's traversal experience during the data acquisition phase, coined traversability masks. The training procedure of the entire model is, therefore, free from manual annotation. In our experiment, we show that the proposed recognition framework is capable of distinguishing traversable plants more accurately than a conventional semantic segmentation with traversable plant and non-traversable plant classes, and an existing image-based traversability estimation method. We also conducted a real-world experiment and confirmed that the robot with the proposed recognition method successfully navigated in plant-rich environments.

Explaining face representation in the primate brain using different computational models

Understanding how the brain represents the identity of complex objects is a central challenge of visual neuroscience. The principles governing object processing have been extensively studied in the macaque face patch system, a sub-network of inferotemporal (IT) cortex specialized for face processing. A previous study reported that single face patch neurons encode axes of a generative model called the "active appearance" model, which transforms 50D feature vectors separately representing facial shape and facial texture into facial images. However, a systematic investigation comparing this model to other computational models, especially convolutional neural network models that have shown success in explaining neural responses in the ventral visual stream, has been lacking. Here, we recorded responses of cells in the most anterior face patch anterior medial (AM) to a large set of real face images and compared a large number of models for explaining neural responses. We found that the active appearance model better explained responses than any other model except CORnet-Z, a feedforward deep neural network trained on general object classification to classify non-face images, whose performance it tied on some face image sets and exceeded on others. Surprisingly, deep neural networks trained specifically on facial identification did not explain neural responses well. A major reason is that units in the network, unlike neurons, are less modulated by face-related factors unrelated to facial identification, such as illumination.

The nearest polyhedral convex conic regions for high-dimensional classification

In the nearest-convex-model type classifiers, each class in the training set is approximated with a convexclass model, and a test sample is assigned to a class based on the shortest distance from the test sample to these classmodels. In this paper, we propose new methods for approximating the distances from test samples to the convex regionsspanned by training samples of classes. To this end, we approximate each class region with a polyhedral convex conicregion by utilizing polyhedral conic functions (PCFs) and its extension, extended PCFs. Then, we derive the necessary formulations for computing the distances from test samples to these new models. We tested the proposed methodson different high-dimensional classification tasks including face, digit, and generic object classification as well as onsome lower-dimensional classification problems. The experimental results on different datasets show that the proposedclassifiers achieve either the best or comparable results on high-dimensional classification problems compared to othernearest-convex-model classifiers, which shows the superiority of the proposed methods.

NAS-FAS: Static-Dynamic Central Difference Network Search for Face Anti-Spoofing.

Face anti-spoofing (FAS) plays a vital role in securing face recognition systems. Existing methods heavily rely on the expert-designed networks, which may lead to a sub-optimal solution for FAS task. Here we propose the first FAS method based on neural architecture search (NAS), called NAS-FAS, to discover the well-suited task-aware networks. Unlike previous NAS works mainly focus on developing efficient search strategies in generic object classification, we pay more attention to study the search spaces for FAS task. The challenges of utilizing NAS for FAS are in two folds: the networks searched on 1) a specific acquisition condition might perform poorly in unseen conditions, and 2) particular spoofing attacks might generalize badly for unseen attacks. To overcome these two issues, we develop a novel search space consisting of central difference convolution and pooling operators. Moreover, an efficient static-dynamic representation is exploited for fully mining the FAS-aware spatio-temporal discrepancy. Besides, we propose Domain/Type-aware Meta-NAS, which leverages cross-domain/type knowledge for robust searching. Finally, in order to evaluate the NAS transferability for cross datasets and unknown attack types, we release a large-scale 3D mask dataset, namely CASIA-SURF 3DMask, for supporting the new 'cross-dataset cross-type' testing protocol. Experiments demonstrate that the proposed NAS-FAS achieves state-of-the-art performance on nine FAS benchmark datasets with four testing protocols.

Deep Learning based MURA Defect Detection

INTRODUCTION: MURA defects in LED/LCD panels are one of the most challenging defects for Automatic Defect Classification and Localization (ADC) due to their extremely low contrast when compared with the background. Manual detection is subjective, error prone, very tedious and time consuming. Even when the type of MURA defects can be ascertained manually, the exact bounding box for defect is hard to determine. Various heuristic based image processing techniques have been applied giving sub-optimal accuracy over generic datasets. OBJECTIVES: The primary objective of this paper is to check whether the state of the art DL (Deep Learning) network for general object classification and localization (MSCOCO PASCAL VOC etc.) can be applied successfully for MURA Defect Classification and Localization. METHODS: In this paper we present a single DL pipeline for classification and localization which for the first time is applied for MURA defects. Naive DL network - Single Shot multi-box Detector (SSD, pre-trained on ImageNet) was not sufficient to give a good F1 score because of the nature of the defect. Accuracy improved a little after applying various DL specific optimization methods such as loss function optimization, network optimization etc. Utilizing the knowledge from MURA domain for data augmentation, like filtering based on image capture wavelength etc. improved the results significantly. RESULTS: Using optimization techniques that are from both DL domain as well as specific to MURA domain, we show improvement in the accuracy of the base DL pipeline from ~30% to ~80%. Minimum heuristics were used to define the pipeline so that it can easily adapt to any new MURA dataset. The paper shows the importance of domain specific preprocessing steps for the designed network in case of MURA defects. CONCLUSION: Using DL, MURA classification and localization had not been tried before. For the first time we demonstrated results for both classification and localization of MURA defects using state-of-the-art DL network with F1~80%. We also conclude that state-of-the-art network for general object detection can be reused with the help of Transfer Learning (TL) concept and fine-tuned with MURA domain specific optimizations mentioned in paper for optimal performances in MURA domain.

Feature visualization in comic artist classification using deep neural networks

Deep neural networks have become a standard framework for image analytics. Besides the traditional applications, such as object classification and detection, the latest studies have started to expand the scope of the applications to include artworks. However, popular art forms, such as comics, have been ignored in this trend. This study investigates visual features for comic classification using deep neural networks. An effective input format for comic classification is first defined, and a convolutional neural network is used to classify comic images into eight different artist categories. Using a publicly available dataset, the trained model obtains a mean F1 score of 84% for the classification. A feature visualization technique is also applied to the trained classifier, to verify the internal visual characteristics that succeed in classification. The experimental result shows that the visualized features are significantly different from those of general object classification. This work represents one of the first attempts to examine the visual characteristics of comics using feature visualization, in terms of comic author classification with deep neural networks.

(224) An Open-Source Deep Neural Network to Assess Spontaneous Pain in Freely-Moving Mice using Facial Expressions

There is a dire need for new non-opioid analgesics. However, efforts to develop new pain medicines have thus far met with limited success. This failure is partly due to an overreliance on evoked pain measures in preclinical models. Most preclinical models do not measure spontaneous pain—the main symptom of chronic pain in humans. Here we show our progress towards developing and validating a fully-automated and user-friendly spontaneous pain measurement tool by training a machine learning model to detect and quantify facial grimacing in mice. The original Mouse Grimace Scale (MGS) was developed to train researchers to detect pain in freely moving mice by quantifying characteristic facial expressions, or “grimacing”. To increase the rate of adoption for this spontaneous pain measure, we recently adapted and trained a general object classification convolutional neural network (Google's Inception v1.0TM) to detect and classify the presence or absence of grimace in still images of freely-moving white mice. We call this recently published model the automated Mouse Grimace Scale, aMGS. Here we show our continuing efforts to further improve our model using new convolutional neural architecture to detect grimacing in mice of various coat colors and across different laboratory settings. Unlike our first aMGS program, our newest model can track mouse faces of various colors using continuous video input, rather than relying on still images to determine pain status. We also show the results of our latest validation experiments, including grimace assessment in mice following laparotomy, a common post-surgical pain assay. The rapidity and utility of our proposed automated pain classifier has the potential to greatly increase the adoption of the mouse grimace as a reproducible and clinically-relevant pain measure, and allow researchers to objectively and rapidly assess tonic and chronic pain states in a variety of mouse models using a measure of spontaneous pain.

Skeleton-Based 3D Object Retrieval Using Retina-Like Feature Descriptor

Skeleton-based 3D object retrieval is a very efficient method to query the sketch databases in numerous applications. However, few skeleton images are found so far in existing sketch benchmarks. In this paper, we provide an initial benchmark dataset consisting of skeleton sketches, including hand-drawn skeletons and skeletons extracted from 3D objects, and both of them are used to form a generic object class. Then we present a method for skeleton-based 3D object retrieval using a retina-like feature descriptor (S3DOR-RFD) based on the structural property of the human retina for processing complex visual information in a very efficient way. As part of the S3DOR-RFD algorithm, we combine artificial bee colony (ABC) in support vector machine (SVM) so as to improve the performance with automatic parameter selection, where one can make full use of the advantages of ABC and SVM to further improve the accuracy rate of 3D object retrieval. Experimental results indicate that skeleton sketches can be automatically distinguished from perspective sketches, and that the proposed S3DOR-RFD method works efficiently for selected object classes.

Sixty-one thousand recent planktonic foraminifera from the Atlantic Ocean

Marine microfossils record the environmental, ecological, and evolutionary dynamics of past oceans in temporally expanded sedimentary archives. Rapid imaging approaches provide a means of exploiting the primary advantage of this archive, the vast number of fossils, for evolution and ecology. Here we provide the first large scale image and 2D and 3D shape dataset of modern planktonic foraminifera, a major microfossil group, from 34 Atlantic Ocean sediment samples. Information on more than 124,000 objects is provided, including general object classification for 4/5ths of the dataset (~ 99,000 objects). Of the ~ 99,000 classifications provided, more than 61,000 are complete or damaged planktonic foraminifera. Objects also include benthic foraminifera, ostracods, pteropods, spicules, and planktonic foraminifera test fragments, among others. This dataset is the first major microfossil output of a new high-throughput imaging method (AutoMorph) developed to extract 2D and 3D data from photographic images of fossils. Our sample preparation and imaging techniques are described in detail. The data provided here comprises the most extensive publically available archive of planktonic foraminiferal morphology and morphological variation to date.

Deep CNNs With Spatially Weighted Pooling for Fine-Grained Car Recognition

Fine-grained car recognition aims to recognize the category information of a car, such as car make, car model, or even the year of manufacture. A number of recent studies have shown that a deep convolutional neural network (DCNN) trained on a large-scale data set can achieve impressive results at a range of generic object classification tasks. In this paper, we propose a spatially weighted pooling (SWP) strategy, which considerably improves the robustness and effectiveness of the feature representation of most dominant DCNNs. More specifically, the SWP is a novel pooling layer, which contains a predefined number of spatially weighted masks or pooling channels. The SWP pools the extracted features of DCNNs with the guidance of its learnt masks, which measures the importance of the spatial units in terms of discriminative power. As the existing methods that apply uniform grid pooling on the convolutional feature maps of DCNNs, the proposed method can extract the convolutional features and generate the pooling channels from a single DCNN. Thus minimal modification is needed in terms of implementation. Moreover, the parameters of the SWP layer can be learned in the end-to-end training process of the DCNN. By applying our method to several fine-grained car recognition data sets, we demonstrate that the proposed method can achieve better performance than recent approaches in the literature. We advance the state-of-the-art results by improving the accuracy from 92.6% to 93.1% on the Stanford Cars-196 data set and 91.2% to 97.6% on the recent CompCars data set. We have also tested the proposed method on two additional large-scale data sets with impressive results observed.

Photo Aesthetics Analysis via DCNN Feature Encoding

We propose an automatic framework for quality assessment of a photograph as well as analysis of its aesthetic attributes. In contrast to the previous methods that rely on manually designed features to account for photo aesthetics, our method automatically extracts such features using a pretrained deep convolutional neural network (DCNN). To make the DCNN-extracted features more suited to our target tasks of photo quality assessment and aesthetic attribute analysis, we propose a novel feature encoding scheme, which supports vector machines-driven sparse restricted Boltzmann machines, which enhances sparseness of features and discrimination between target classes. Experimental results show that our method outperforms the current state-of-the-art methods in automatic photo quality assessment, and gives aesthetic attribute ratings that can be used for photo editing. We demonstrate that our feature encoding scheme can also be applied to general object classification task to achieve performance gains.

An Ensemble of Invariant Features for Person Reidentification

This paper proposes an ensemble of invariant features (EIFs), which can properly handle the variations of color difference and human poses/viewpoints for matching pedestrian images observed in different cameras with nonoverlapping field of views. Our proposed method is a direct reidentification (re-id) method, which requires no prior domain learning based on prelabeled corresponding training data. The novel features consist of the holistic and region-based features. The holistic features are extracted by using a publicly available pretrained deep convolutional neural network used in generic object classification. In contrast, the region-based features are extracted based on our proposed two-way Gaussian mixture model fitting, which overcomes the self-occlusion and pose variations. To make a better generalization during recognizing identities without additional learning, the ensemble scheme aggregates all the feature distances using the similarity normalization. The proposed framework achieves robustness against partial occlusion, pose, and viewpoint changes. Moreover, the evaluation results show that our method outperforms the state-of-the-art direct re-id methods on the challenging benchmark viewpoint invariant pedestrian recognition and 3D people surveillance data sets.

Plant species classification using flower images—A comparative study of local feature representations

Steady improvements of image description methods induced a growing interest in image-based plant species classification, a task vital to the study of biodiversity and ecological sensitivity. Various techniques have been proposed for general object classification over the past years and several of them have already been studied for plant species classification. However, results of these studies are selective in the evaluated steps of a classification pipeline, in the utilized datasets for evaluation, and in the compared baseline methods. No study is available that evaluates the main competing methods for building an image representation on the same datasets allowing for generalized findings regarding flower-based plant species classification. The aim of this paper is to comparatively evaluate methods, method combinations, and their parameters towards classification accuracy. The investigated methods span from detection, extraction, fusion, pooling, to encoding of local features for quantifying shape and color information of flower images. We selected the flower image datasets Oxford Flower 17 and Oxford Flower 102 as well as our own Jena Flower 30 dataset for our experiments. Findings show large differences among the various studied techniques and that their wisely chosen orchestration allows for high accuracies in species classification. We further found that true local feature detectors in combination with advanced encoding methods yield higher classification results at lower computational costs compared to commonly used dense sampling and spatial pooling methods. Color was found to be an indispensable feature for high classification results, especially while preserving spatial correspondence to gray-level features. In result, our study provides a comprehensive overview of competing techniques and the implications of their main parameters for flower-based plant species classification.

A survey on object detection in optical remote sensing images

Object detection in optical remote sensing images, being a fundamental but challenging problem in the field of aerial and satellite image analysis, plays an important role for a wide range of applications and is receiving significant attention in recent years. While enormous methods exist, a deep review of the literature concerning generic object detection is still lacking. This paper aims to provide a review of the recent progress in this field. Different from several previously published surveys that focus on a specific object class such as building and road, we concentrate on more generic object categories including, but are not limited to, road, building, tree, vehicle, ship, airport, urban-area. Covering about 270 publications we survey (1) template matching-based object detection methods, (2) knowledge-based object detection methods, (3) object-based image analysis (OBIA)-based object detection methods, (4) machine learning-based object detection methods, and (5) five publicly available datasets and three standard evaluation metrics. We also discuss the challenges of current studies and propose two promising research directions, namely deep learning-based feature representation and weakly supervised learning-based geospatial object detection. It is our hope that this survey will be beneficial for the researchers to have better understanding of this research field.

Automatic classification for field crop insects via multiple-task sparse representation and multiple-kernel learning

Classification of insect species of field crops such as corn, soybeans, wheat, and canola is more difficult than the generic object classification because of high appearance similarity among insect species. To improve the classification accuracy, we develop an insect recognition system using advanced multiple-task sparse representation and multiple-kernel learning (MKL) techniques. As different features of insect images contribute differently to the classification of insect species, the multiple-task sparse representation technique can combine multiple features of insect species to enhance the recognition performance. Instead of using hand-crafted descriptors, our idea of sparse-coding histograms is adopted to represent insect images so that raw features (e.g., color, shape, and texture) can be well quantified. Furthermore, the MKL method is proposed to fuse multiple features effectively. The proposed learning model can be optimized efficiently by jointly optimizing the kernel weights. Experimental results on 24 common pest species of field crops show that our proposed method performs well on the classification of insect species, and outperforms the state-of-the-art methods of the generic insect categorization.