Structural crack detection using deep convolutional neural networks

With the development of advanced information and intelligence technologies, precision agriculture has become an effective solution to monitor and prevent crop pests and diseases. However, pest and disease recognition in precision agriculture applications is essentially the fine-grained image classification task, which aims to learn effective discriminative features that can identify the subtle differences among similar visual samples. It is still challenging to solve for existing standard models troubled by oversized parameters and low accuracy performance. Therefore, in this paper, we propose a feature-enhanced attention neural network (Fe-Net) to handle the fine-grained image recognition of crop pests and diseases in innovative agronomy practices. This model is established based on an improved CSP-stage backbone network, which offers massive channel-shuffled features in various dimensions and sizes. Then, a spatial feature-enhanced attention module is added to exploit the spatial interrelationship between different semantic regions. Finally, the proposed Fe-Net employs a higher-order pooling module to mine more highly representative features by computing the square root of the covariance matrix of elements. The whole architecture is efficiently trained in an end-to-end way without additional manipulation. With comparative experiments on the CropDP-181 Dataset, the proposed Fe-Net achieves Top-1 Accuracy up to 85.29% with an average recognition time of only 71 ms, outperforming other existing methods. More experimental evidence demonstrates that our approach obtains a balance between the model’s performance and parameters, which is suitable for its practical deployment in precision agriculture art applications.

/pdf/a-spatial-feature-enhanced-attention-neural-network-with-4ogxn80u.pdf

A Spatial Feature-Enhanced Attention Neural Network with High-Order Pooling Representation for Application in Pest and Disease Recognition

A key notion in numerous data mining and machine learning (ML) algorithms says that the training data and testing data are essentially in the similar feature space and also have the alike probability distribution function (PDF). Though, in several real-life applications, this theory might not retain true. There are issues where training data is costly or tough to gather. Thus, there is a necessity to build high-performance classifiers, trained using more commonly found data from distinct domains. This methodology is stated as transfer learning (TL). TL is usually beneficial when enough data is not available in the target domain but the large dataset is available in source domain. This survey paper explains transfer learning along with its categorization and provides examples and perspective related to transfer learning. Negative transfer learning is also discussed in detail along with its effects on the accomplishment of learning in target domain.

Transfer Learning: Survey and Classification

Component-based face recognition under transfer learning for forensic applications

Biometrics is a set of highly automated methods used for recognition purposes including analyzing physical traits of people and statistically measuring them. In forensic applications, fingerprint and face images are mostly used for recognition. There is a need during a criminal investigation to find out a face image of a person from its fingerprint. The proposed method able to identify the face of a person using an associated fingerprint. In this paper, Bregman divergence regularization is used to learn and optimize transferring subspace. This method first gleans knowledge from samples that are meant for training and transfer it to the testing samples. This regularization helps to minimize the probability distribution differences between two different domains. It helps to find a common subspace that boosts the performance of independent and identically distributed (i.i.d.) condition of samples. Practically the samples violate this i.i.d. condition. However, it will help to identify the correct suspect.

Transfer learning for face recognition using fingerprint biometrics

Face and facial components are the most significant pieces of evidence used in forensic applications. Component-based recognition and age-invariant face recognition are the key factors in forensic face recognition. A major challenge during a forensic investigation is the association of facial components with a relevant face and this paper addresses this concern using the proposed transfer learning methodology. Research has been carried out treating facial components and the face as individual and separate entities. However, they share common auxiliary information which is exploited to bring the face and facial components in a common subspace in the proposed method. This facilitates transfer of knowledge gained from a face and classification of facial components of a person using the same subspace. Stability and invariant features of the facial components catalyze a solution for the problem of recognition. Convolutional neural network (CNN) is used to extract the features of the biometrics and then they are given to a regularization framework for reducing the probability distribution difference between them. The system is trained using the face and facial components and then tested using either the full-face or individual facial components. For the transfer, the proposed Fisher linear discriminant analysis and locality preserving projection, a convolutional neural network-based algorithm gives 91% and 90% accuracy, respectively, which outperforms the Histogram of Gradient and Gabor methods for predicting an association.

https://link.springer.com/content/pdf/10.1007/s42979-020-00280-2.pdf

Association of Face and Facial Components Based on CNN and Transfer Subspace Learning for Forensics Applications

Transfer learning has gained more attention recently by utilizing knowledge acquired from one domain to advance a learning performance in another domain. Existing homogeneous transfer learning methods have progressed to a point where feature spaces are common in training and testing domains. However, heterogeneous transfer learning is still in its nascent stage where features of training and testing domains are different. Taking this into account, Bregman Divergence Regularization is used to minimize the probability distribution difference between training and testing domains and to take them together to a shared subspace. To discriminate data within individual domains, a projection matrix is obtained using Fisher Linear Discriminant Analysis subspace learning algorithm. Experimentation is performed on two efficiently used biometrics: the face and fingerprint. Two types of cross-domain settings are used: (1) Face + Finger2Finger where training samples come from face (labeled samples) and fingerprint (unlabeled samples) data sets, while testing is performed on a fingerprint dataset. (2) Finger + Face2Face where training samples come from fingerprint (labeled samples) and face (unlabeled samples) data sets while testing is performed on a face dataset. This paper proposes a cross domain association between face and fingerprint that finds utility in forensic applications.

Cross domain association using transfer subspace learning

Different biometrics studies related to face, fingerprint, hand geometry, palmprint, and iris has been taken place in recent past. In this paper two biometrics, face and fingerprint are considered and associated with each other using transfer subspace learning. Transfer subspace learning can share common subspace even if domains are different by minimizing the distance between their probability distribution. Locality preserving projections is used as objective function to discriminate between source domains. This works linearly and preserves the local geometry of the structure by obtaining the subspace spanned by smallest eigenvectors of local covariance matrix. Proposed framework will be very useful in the forensic applications. Though the accuracy of this system is not very high it will help to separate the most probable suspects.

Rupali Sandip Kute

Papers

Component-based face recognition under transfer learning for forensic applications

Transfer learning for face recognition using fingerprint biometrics

Association of Face and Facial Components Based on CNN and Transfer Subspace Learning for Forensics Applications

Cross domain association using transfer subspace learning

Biometric association using transfer subspace learning