Nondestructive detection methods, based on vibrational spectroscopy, are vitally important in a wide range of applications including industrial chemistry, pharmacy and national defense. Recently, deep learning has been introduced into vibrational spectroscopy showing great potential. Different from images, text, etc. that offer large labeled data sets, vibrational spectroscopic data is very limited, which requires novel concepts beyond transfer and meta learning. To tackle this, we propose a task-enhanced augmentation network (TeaNet). The key component of TeaNet is a reconstruction module that inputs randomly masked spectra and outputs reconstructed samples that are similar to the original ones, but include additional variations learned from the domain. These augmented samples are used to train the classification model. The reconstruction and prediction parts are trained simultaneously, end-to-end with back-propagation. Results on both synthetic and real-world datasets verified the superiority of the proposed method. In the most difficult synthetic scenarios TeaNet outperformed CNN by 17%. We visualized and analysed the neuron responses of TeaNet and CNN, and found that TeaNet's ability to identify discriminant wavenumbers was excellent compared to CNN. Our approach is general and can be easily adapted to other domains, offering a solution to more accurate and interpretable few-shot learning.

Cross-modality recognition has many important applications in science, law enforcement and entertainment. Popular methods to bridge the modality gap include reducing the distributional differences of representations of different modalities, learning indistinguishable representations or explicit modality transfer. The first two approaches suffer from the loss of discriminant information while removing the modality-specific variations. The third one heavily relies on the successful modality transfer, could face catastrophic performance drop when explicit modality transfers are not possible or difficult. To tackle this problem, we proposed a compact encoder-decoder neural module (cmUNet) to learn modality-agnostic representations while retaining identity-related information. This is achieved through cross-modality transformation and in-modality reconstruction, enhanced by an adversarial/perceptual loss which encourages indistinguishability of representations in the original sample space. For cross-modality matching, we propose MarrNet where cmUNet is connected to a standard feature extraction network which takes as inputs the modality-agnostic representations and outputs similarity scores for matching. We validated our method on five challenging tasks, namely Raman-infrared spectrum matching, cross-modality person re-identification and heterogeneous (photo-sketch, visible-near infrared and visible-thermal) face recognition, where MarrNet showed superior performance compared to state-of-the-art methods. Furthermore, it is observed that a cross-modality matching method could be biased to extract discriminant information from partial or even wrong regions, due to incompetence of dealing with modality gaps, which subsequently leads to poor generalization. We show that robustness to occlusions can be an indicator of whether a method can well bridge the modality gap. This, to our knowledge, has been largely neglected in the previous works. Our experiments demonstrated that MarrNet exhibited excellent robustness against disguises and occlusions, and outperformed existing methods with a large margin (>10%). The proposed cmUNet is a meta-approach and can be used as a building block for various applications.

Recently, many methods based on hand-designed convolutional neural networks (CNNs) have achieved promising results in automatic retinal vessel segmentation. However, these CNNs remain constrained in capturing retinal vessels in complex fundus images. To improve their segmentation performance, these CNNs tend to have many parameters, which may lead to overfitting and high computational complexity. Moreover, the manual design of competitive CNNs is time-consuming and requires extensive empirical knowledge. Herein, a novel automated design method, called Genetic U-Net, is proposed to generate a U-shaped CNN that can achieve better retinal vessel segmentation but with fewer architecture-based parameters, thereby addressing the above issues. First, we devised a condensed but flexible search space based on a U-shaped encoder-decoder. Then, we used an improved genetic algorithm to identify better-performing architectures in the search space and investigated the possibility of finding a superior network architecture with fewer parameters. The experimental results show that the architecture obtained using the proposed method offered a superior performance with less than 1% of the number of the original U-Net parameters in particular and with significantly fewer parameters than other state-of-the-art models. Furthermore, through in-depth investigation of the experimental results, several effective operations and patterns of networks to generate superior retinal vessel segmentations were identified. The codes of this work are available at https://github.com/96jhwei/Genetic-U-Net.

Detecting the shape of the non-rigid molten metal during welding, so-called weld pool visual sensing, is one of the central tasks for automating arc welding processes. It is challenging due to the strong interference of the high-intensity arc light and spatters as well as the lack of robust approaches to detect and represent the shape of the nonrigid weld pool. We propose a solution using active contours including an prior for the weld pool boundary composition. Also, we apply Adaboost to select a small set of features that captures the relevant information. The proposed method is applied to weld pool tracking and the presented results verified its feasibility.

In this paper, we propose a methodology to improve the performance of the standard Differential Evolution (DE) in constraint optimization applications, in terms of accelerating its search speed, and improving the success rate. One critical mechanism embedded in the approach is applying Stochastic Ranking (SR) to rank the whole population of individuals with both objective value and constraint violation to be compared. The ranked population is then in a better shape to provide useful information e.g. direction to guide the search process. The strength of utilizing the directional information can be further controlled by a parameter - population partitioning factor, which is adjusted according to the evolution stage and generations. Because the adaptive adjustment of the parameter is predefined and does not need user input, the resulting algorithm is free of definition of this extra parameter and easier to implement. The performance of the proposed approach, which we call SRaDE (Stochastic Ranking based Adaptive Differential Evolution) is investigated and compared with standard DE. The experimental results show that SRDE significantly outperforms, or at least is comparable with standard DE in all the tested benchmark functions. We also conducted an experiment to compare SRaDE with SRDE - a variant of Stochastic Ranking based Differential Evolution without adaptive adjustment of the population partitioning factor. Experimental results show that SRaDE can also achieve improved performance over SRDE.

In the recent decades much research has been performed in order to allow better control of arc welding processes, but the success has been limited, and the vast majority of the industrial structural welding work is therefore still being made manually. Closed-loop and nearly-closed-loop control of the processes requires the extraction of characteristic parameters of the welding groove close to the molten pool, i.e. in an environment dominated by the very intense light emission from the welding arc. The typical industrial solution today is a laser-scanner containing a camera as well as a laser source illuminating the groove by a light curtain and thus allowing details of the groove geometry to be extracted by triangulation. This solution is relatively expensive and must act several centimetres ahead of the molten pool. In addition laser-scanners often show problems when dealing with shiny surfaces. It is highly desirable to extract groove features closer to the arc and thus facilitate for a nearly-closed-loop control situation. On the other hand, for performing seam tracking and nearly-closed-loop control it is not necessary to obtain very detailed information about the molten pool area as long as some important features are obtained, e.g. the groove position and gap width. To obtain these features without external illumination, a new image analysis scheme based on active contour models was proposed and verified by experimental results.