Image caption generator using deep learning architecture with LSTM and their comparative analysis

Steering angle prediction plays a crucial role in the control of Autonomous Vehicles (AVs) and has garnered the interest of researchers, manufacturers, and insurance companies within the automotive industry. Various Deep Learning (DL) architectures have been utilized to forecast the steering angle of AVs across different scenarios. An examination focusing on steering angle prediction through deep learning algorithms can assist seasoned researchers in pinpointing areas that necessitate further advancement. Additionally, newcomers in the field can utilize this study as a foundation. This paper offers an extensive analysis of the recent progress achieved in DL architectures concerning the steering angle prediction of AVs. A comprehensive taxonomy outlining the application of DL in steering angle prediction of AVs has been developed. The survey provides a succinct synthesis, summary, and analysis of research findings. It is evident that Convolutional Neural Network (CNN) is the preferred choice among researchers for predicting the steering angle of autonomous vehicles, compared to other DL architectures. Furthermore, existing research challenges have been identified. The primary obstacle encountered in DL-based steering angle prediction of AVs is the scarcity of real-world datasets, leading researchers to heavily rely on data generated from simulated environments. Lastly, potential alternative approaches to address the identified research challenges have been suggested, indicating promising avenues for future research endeavours.

Deep learning based classification of breast tumors with shear-wave elastography

Detection of smoke from videos captured by surveillance cameras in outdoor environments is one of the useful outcome of Internet of Things (IoT) applications. The potential benefit increases when deep learning (DL) architectures are involved. However, an inherent difficulty is to detect smoke while natural events like fog exists. The effectiveness of color spaces in detection performance has not yet fully evaluated in those architectures. Moreover, the energy and memory requirements of DL architectures may not be applicable for handling IoT implementation demands. Therefore, in this work, a DL architecture with a suitable color space model, applicable for IoT implementations is proposed to detect smoke from videos in foggy environment. By collecting several videos including smoke samples, the performance comparison of popular and the state-of-the-art DL architectures denoted the outperforming result according to both accuracy and memory usage.

Recent advancements in technology have made available a variety of image capturing devices, ranging from handheld mobiles to space-grade rovers. This has generated a tremendous visual data, which has made a necessity to organize and understand this visual data. Thus, there is a need to caption thousands of such images. This has resulted in tremendous research in computer vision and deep learning. Inspired by such recent works, we present an image caption generating system that uses convolutional neural network (CNN) for extracting the feature embedding of an image and feed that as an input to long short-term memory cells that generates a caption. We are using two pre-trained CNN models on ImageNet, VGG16 and ResNet-101. Both the models were tested and compared on Flickr8K dataset. Experimental results on this dataset demonstrate that the proposed architecture is as efficient as the state-of-the-art multi-label classification models. Experimental results on public benchmark dataset demonstrate that the proposed architecture performs as efficiently as the state-of-the-art image captioning model.

This study proposes a compact deep learning (DL) architecture and a highly parallelized computing hardware platform to reconstruct the blood flow index (BFi) in diffuse correlation spectroscopy (DCS). We leveraged a rigorous analytical model to generate autocorrelation functions (ACFs) to train the DL network. We assessed the accuracy of the proposed DL using simulated and milk phantom data. Compared to convolutional neural networks (CNN), our lightweight DL architecture achieves 66.7% and 18.5% improvement in MSE for BFi and the coherence factor β, using synthetic data evaluation. The accuracy of rBFi over different algorithms was also investigated. We further simplified the DL computing primitives using subtraction for feature extraction, considering further hardware implementation. We extensively explored computing parallelism and fixed-point quantization within the DL architecture. With the DL model's compact size, we employed unrolling and pipelining optimizations for computation-intensive for-loops in the DL model while storing all learned parameters in on-chip BRAMs. We also achieved pixel-wise parallelism, enabling simultaneous, real-time processing of 10 and 15 autocorrelation functions on Zynq-7000 and Zynq-UltraScale+ field programmable gate array (FPGA), respectively. Unlike existing FPGA accelerators that produce BFi and the β from autocorrelation functions on standalone hardware, our approach is an encapsulated, end-to-end on-chip conversion process from intensity photon data to the temporal intensity ACF and subsequently reconstructing BFi and β. This hardware platform achieves an on-chip solution to replace post-processing and miniaturize modern DCS systems that use single-photon cameras. We also comprehensively compared the computational efficiency of our FPGA accelerator to CPU and GPU solutions.

AI-based emotion recognition has emerged as a critical component of affect-aware educational technologies, particularly in online, large-scale, and technology-mediated learning environments where direct observation of learners’ emotional states is limited. By leveraging facial expressions, speech, physiological signals, and interaction data, emotion-aware systems aim to support adaptive instruction, learner engagement, and instructional decision-making. Despite rapid technical progress, the integration of emotion recognition into authentic educational practice remains constrained by dataset limitations, cultural bias, weak pedagogical alignment, and unresolved ethical concerns. This article presents a critical and theory-informed review of recent advances in AI-based emotion recognition in education, with particular emphasis on deep learning architectures and multimodal fusion approaches, while highlighting persistent challenges and future research directions.Purpose:This study aims to critically examine the current state of AI-based emotion recognition in educational contexts by analyzing dominant modeling paradigms, sensing modalities, application domains, and ethical implications, in order to identify key gaps between technical development and pedagogical practice.Methods/Study design/approach:The study adopts a narrative literature review approach grounded in educational and affective computing theories. Representative and influential studies were analyzed across facial, speech, physiological, and multimodal emotion recognition systems, with emphasis on deep learning, attention-based, and transformer architectures. The review synthesizes findings related to system performance, educational applications, dataset characteristics, and ethical considerations.Result/Findings:The findings indicate that multimodal emotion recognition systems generally outperform unimodal approaches in detecting learning-relevant affective states such as engagement, frustration, and confusion. Advances in deep learning, particularly attention-based and transformer models, improve temporal modeling and robustness. However, most systems rely on non-educational datasets, demonstrate limited cultural generalizability, and prioritize recognition accuracy over pedagogical interpretability and long-term educational impact. Ethical challenges related to privacy, consent, bias, and emotional surveillance remain insufficiently addressed.Novelty/Originality/Value:This study provides an integrative and pedagogically grounded synthesis that bridges affective computing research with educational theory and governance perspectives. By reframing AI-based emotion recognition as a decision-support mechanism rather than an automated control system, the article offers a conceptual foundation for responsible, culturally inclusive, and ethically aligned deployment of emotion-aware AI in education, serving as a reference for future interdisciplinary research and policy development.

Natural language processing (NLP) techniques have evolved substantially in recent years, with transformer-based architectures emerging as powerful tools for textual analysis across various fields. This paper investigates the underexplored area of Arabic legal text classification through the implementation of transformer-based language models integrated with deep learning (DL) architectures. Even though transformer models are widely used in legal NLP for high-resource languages, their application to Arabic legal texts remains limited due to linguistic complexities and dataset scarcity. We introduce a classification model for Arabic legal text and present a new dataset of over 27,000 Moroccan legal cases with seven categories. Our methodology implements a segment-based approach to overcome the token limitations of transformer models, processing document segments individually before aggregating embeddings to create complete document representations. Four deep learning architectures are then used to classify those documents. Experimental evaluations demonstrate that MARBERTv2 combined with Gated Recurrent Unit (GRU) architecture achieves superior performance with 95.01% accuracy and Matthews Correlation Coefficient (MCC) of 0.9090.

The aims of this work are (1) to explore deep learning (DL) architectures, spectroscopic input types, and learning designs toward optimal quantification in MR spectroscopy of simulated pathological spectra; and (2) to demonstrate accuracy and precision of DL predictions in view of inherent bias toward the training distribution. Simulated 1D spectra and 2D spectrograms that mimic an extensive range of pathological in vivo conditions are used to train and test 24 different DL architectures. Active learning through altered training and testing data distributions is probed to optimize quantification performance. Ensembles of networks are explored to improve DL robustness and reduce the variance of estimates. A set of scores compares performances of DL predictions and traditional model fitting (MF). Ensembles of heterogeneous networks that combine 1D frequency-domain and 2D time-frequency domain spectrograms as input perform best. Dataset augmentation with active learning can improve performance, but gains are limited. MF is more accurate, although DL appears to be more precise at low SNR. However, this overall improved precision originates from a strong bias for cases with high uncertainty toward the dataset the network has been trained with, tending toward its average value. MF mostly performs better compared to the faster DL approach. Potential intrinsic biases on training sets are dangerous in a clinical context that requires the algorithm to be unbiased to outliers (i.e., pathological data). Active learning and ensemble of networks are good strategies to improve prediction performances. However, data quality (sufficient SNR) has proven as a bottleneck for adequate unbiased performance-like in the case of MF.

This study compared popular Deep Learning (DL) architectures to classify machining surface roughness using sound and force data. The DL architectures considered in this study include Multi-Layer Perceptron (MLP), Convolution Neural Network (CNN), Long Short-Term Memory (LSTM), and transformer. The classification was performed on the sound and force data generated during machining aluminum sheets for different levels of spindle speed, feed rate, depth of cut, and end-mill diameter, and it was trained on 30 s machining data (10–40 s) of the machining experiments. Since a raw audio waveform is seldom used in DL models, Mel-Spectrogram and Mel Frequency Cepstral Coefficients (MFCCs) audio feature extraction techniques were used in the DL models. The results of DL models were compared for the training–validation accuracy, training epochs, and training parameters of each model. Although the roughness classification by all the DL models was satisfactory (except for CNN with Mel-Spectrogram), the transformer-based modes had the highest training (>96%) and validation accuracies (≈90%). The CNN model with Mel-Spectrogram exhibited the worst training and inference accuracy, which is influenced by limited training data. Confusion matrices were plotted to observe the classification accuracy visually. The confusion matrices showed that the transformer model trained on Mel-Spectrogram and the transformer model trained on MFCCs correctly predicted 366 (or 91.5%) and 371 (or 92.7%) out of 400 test samples. This study also highlights the suitability and superiority of the transformer model for time series sound and force data and over other DL models.

We have developed a magnetic resonance (MR) image-based radiomic biopsy approach for estimation of malignancy grade in parotid gland cancer (PGC). Preoperative T1- and T2-weighted MR images of 39 PGC patients with 20 highand 19 intermediate-/low-malignancy grades were employed. High- versus intermediate-/low-malignancy grades were estimated using MR-radiomic biopsy approaches, i.e. 972 hand-crafted feature and transfer learning of five pre-trained deep learning (DL) architectures (AlexNet, GoogLeNet, VGG-16, ResNet-101, DenseNet-201). The 39 patients were divided into 70% for training datasets and 30% for test datasets. The hand-crafted features were extracted from cancer regions in T1- and T2-weighted MR images. Three features were selected as a radiomic signature by using a least absolute shrinkage and selection operator (LASSO), whose coefficients of three features were used for constructing the radiomic score (Rad-score). The two grade malignancy was estimated by using an optimal cut-off value of Rad-score. On the other hand, last three layers of the DL architectures were replaced with new three layers for the estimation task. The DL architectures were fine-tuned with training datasets and were evaluated with test datasets. The performances of the MR-radiomic biopsy approaches were assessed by using the accuracy and the area under the receiver operating characteristic curve (AUC). The VGG-16 demonstrated the best performance (accuracy=85.4%, AUC=0.906), but the other approaches showed worse performances (Rad-score: 83.3%, 0.830, AlexNet: 84.4%, 0.915, GoogLeNet: 84.9%, 0.884, ResNet-101: 84.9%, 0.918, DenseNet-201: 84.4%, 0.869) than the VGG-16. The VGG-16-based MR-radiomic biopsy could be feasible for the malignancy grade estimation of PGC.

Image Captioning, the process of generating a textual description of an image, has emerged as a hot research due to its practical importance in many domains. It is a challenging task as it uses both Natural Language Processing and Computer Vision related fields to generate the captions. Despite the fact that the literature has reported notable image captioning methodologies, they still lag in accomplishing the substantial performance level for diverse datasets. This paper proposes an image caption generating mechanism based on Optimized Bidirectional Long Short-Term Memory (B-LSTM) model. We propose a variant of Moth Flame Optimization (PMFO), termed here as Proposed Moth Flame Optimization (PMFO), which has logarithmic spiral update based on correlation. The performance of the proposed model is demonstrated on benchmark datasets like Flicker 8 k, Flicker30k, VizWik and COCO datasets using renowned metrics such as CIDEr, BLEU, SPICE and ROUGH. The performance analysis proves that the B-LSTM achieves better performance on caption generation than state-of-the-art methods.

The caption generation is nothing but the generation of textual information from images. For this, objects from images are extracted and classified among predefined classes. The logical objects from the image are extracted and transformed into natural sentences. The recognizing process requires an iterative task that incorporates image recognition as well as machine vision. The process must define relations among objects, persons, and animals and create the textual description of these relations. The paper is about the study of deep learning techniques to discover, identify and produce good captions for a source image. The process of making explanations in the form of sentences for a source image is image captioning, which involves machine vision and natural sentence forming techniques. For these processes, recent models have used deep learning techniques to acquire a great improvement in performance. Second, a more advanced trend is set in utilizing attention-dependent structure for captioning. Recent interpreters use a process of attention for each produced term containing seen term and unseen term. However, these unseen terms are effortlessly detected by considering a model for language in the absence of taking seen indicators, but unseen words could cause and a give bad performance for visual captioning. Taking these issues into consideration, the hierarchy of LSTM [Long–Short-Term Memory] with adaptive attention approach for the creation of captions for images and videos is presented.

In recent years, with the rapid development of artificial intelligence, image caption has gradually attracted the attention of many researchers in the field of artificial intelligence and has become an interesting and arduous task. Image caption, automatically generating natural language descriptions according to the content observed in an image, is an important part of scene understanding, which combines the knowledge of computer vision and natural language processing. The application of image caption is extensive and significant, for example, the realization of human-computer interaction. This paper summarizes the related methods and focuses on the attention mechanism, which plays an important role in computer vision and is recently widely used in image caption generation tasks. Furthermore, the advantages and the shortcomings of these methods are discussed, providing the commonly used datasets and evaluation criteria in this field. Finally, this paper highlights some open challenges in the image caption task.

Enhancing Cross-Linguistic Image Caption Generation with Indian Multilingual Voice Interfaces using Deep Learning Techniques

Image caption generation is a fundamental task to build a bridge between image and its description in text, which is drawing increasing interest in artificial intelligence. Images and textual sentences are viewed as two different carriers of information, which are symmetric and unified in the same content of visual scene. The existing image captioning methods rarely consider generating a final description sentence in a coarse-grained to fine-grained way, which is how humans understand the surrounding scenes; and the generated sentence sometimes only describes coarse-grained image content. Therefore, we propose a coarse-to-fine-grained hierarchical generation method for image captioning, named SDA-CFGHG, to address the two problems above. The core of our SDA-CFGHG method is a sequential dual attention that is used to fuse different grained visual information with sequential means. The advantage of our SDA-CFGHG method is that it can achieve image captioning in a coarse-to-fine-grained way and the generated textual sentence can capture details of the raw image to some degree. Moreover, we validate the impressive performance of our method on benchmark datasets—MS COCO, Flickr—with several popular evaluation metrics—CIDEr, SPICE, METEOR, ROUGE-L, and BLEU.