hajj mobile chat application for piligrims

this contains the research papers

1. Machine Learning-Integrated Usability Evaluation and Monitoring of Human Activities for Individuals With Special Needs During Hajj and Umrah (IEEE Access 2025)

2. Why is anomaly detection in abnormal behavior important Human activity recognition (HAR) is important for the safety and accessibility of individuals with disabilities. This is especially essential during large-scale events like Hajj and Umrah, which attract millions of participants each year. These gatherings pose significant challenges for people with disabilities and affect both their mobility and security.

3. Problem Statement This study introduces a new approach to improve usability and tracking for disabled pilgrims performing the two holy pilgrimages: Hajj and Umrah, which are performed annually by millions of Muslims. These acts of worship are still very challenging for the mobility, security, and accessibility of people with special needs.

4. Method Using clustering, anomaly detection, and predictive modeling, it was intended to enhance the safety and security of sensitive participants. Using the K-Means algorithm and the Elbow Method as the initial indicators to classify the clusters, we reveal four clusters that relate to different human activities that are further visualized through principal component analysis (PCA). Activities are clustered based on the behavioral patterns observed among participants, followed by the performance of anomaly detection

5. Dataset Kaggle human activity recognition with smartphones ( https://www.kaggle.com/datasets/uciml/human-activity-recognition-with-smartphones ) csv format

6. Smart mobile application for real-time monitoring and prediction of pilgrim crowd activities in Hajj and Umrah using wearable sensors (Elsevier 2025)

7. Introduction This study presents a novel Smart Mobile Application for Monitoring and Classification of Pilgrim Activities in Hajj and Umrah Using Wearable Sensors (SMAMCPA-HUWS) method. The main objective of the SMAMCPA-HUWS method is to classify pilgrim activities based on fatigue levels and emotional status using wearable sensors for real-time monitoring.

8. Cont.. The SMAMCPA-HUWS method performs data preprocessing using min-max normalization to standardize input data, ensuring consistent analysis across varied sensor readings. The waterwheel plant algorithm (WWPA) model is utilized for feature subset selection to identify the most relevant features by reducing data dimensionality and enhancing processing efficiency.

9. Cont.. An ensemble of three classifiers, such as bidirectional long short-term memory ( BiLSTM ), bidirectional temporal convolutional network ( BiTCN ), and conditional variational autoencoder (CVAE), is employed for classifying pilgrim activities.

10. Dataset Hajj Crowd Activity Prediction Dataset (https://www.kaggle.com/datasets/sohagalalaldinahmed/hajj-crowd-activity-prediction-dataset) The dataset is saved in .csv format, while the metadata is in . xls format. It can be used for various purposes such as testing hypotheses for fatigue detection, emotional classification, and activity recognition, as well as developing real-time algorithms for detecting Hajj activities. 

11. Results

12. Leveraging Machine Learning for Optimal Pilgrim Crowd Management (electronics 2025)

13. Introduction The Hajj pilgrimage involves high crowd density within a limited time and space, making traditional crowd control methods insufficient for real-time alerts or predictive safety measures.

14. Method This research proposes a machine learning-based system to enhance crowd management by detecting abnormal behavior and forecasting future conditions. The study utilizes the Hajjv2 dataset, which consists of annotated video frames capturing various crowd behaviors across multiple Hajj locations. The methodology involves dataset acquisition, preprocessing, feature engineering, anomaly detection, forecasting, and alert generation.

15. Cont.. After data preprocessing and feature extraction, including crowd density, speed, direction, and object area, two models are employed: The Isolation Forest algorithm for anomaly detection and a Long Short-Term Memory (LSTM) neural network for forecasting crowd behavior. The system integrates the results of both models to issue real-time alerts based on predefined thresholds.

16. Hajjv2_dataset (https://www.kaggle.com/datasets/tarikalafif/hajjv2-dataset) The HAJJv2 dataset is introduced due to the imbalance of training examples in each class and the absence of many annotations and labeling for individuals with abnormal behaviors in the HAJJv1 dataset. The HAJJv2 dataset consists of 18 manually collected videos from the annual Hajj religious event.  CSV Need a Microsoft account for accessing videos

17. Results

18. HYBRID CLASSIFIERS FOR SPATIO-TEMPORAL REAL-TIME ABNORMAL BEHAVIORS DETECTION, TRACKING, AND RECOGNITION IN MASSIVE HAJJ CROWDS (Archive 2022)

19. Introduction Individual abnormal behaviors vary depending on crowd sizes, contexts, and scenes. Challenges such as partial occlusions, blurring, large numbers of abnormal behaviors, and camera viewing occur in large-scale crowds when detecting, tracking, and recognizing individuals with abnormal behaviors.

20. Contributions First, we introduce an annotated and labeled large-scale crowd abnormal behaviors Hajj dataset (HAJJv2). We propose two methods of hybrid CNNs and RFs classifiers to detect, track, and recognize Spatio -temporal abnormal behaviors in small scale and large-scale crowd videos. We evaluate the first proposed method on two common benchmark abnormal behaviors pub lic small-scale crowd video datasets, UMN and UCSD, against the currently published methods. Then, we evaluate the second proposed method on the HAJJv2 dataset and compare it with the previously existing method.

21. Abnormal behaviors in hajjv2 dataset

22. dataset

23. statistics

24. Results

25. Abnormal Activity Recognition from Surveillance Videos Using Convolutional Neural Network (Sensors 2025)

26. Introduction Background and motivation: Every year, millions of Muslims worldwide come to Mecca to perform the Hajj. In order to maintain the security of the pilgrims, the Saudi government has installed about 5000 closed circuit television (CCTV) cameras to monitor crowd activity efficiently. Problem: As a result, these cameras generate an enormous amount of visual data through manual or offline monitoring, requiring numerous human resources for efficient tracking.

27. Cont..

28. contributions A deep learning assistive framework is developed for the efficient recognition of violent activity by using the visual sensor. In order to efficiently utilize our proposed framework, a lightweight CNN object detector is trained on the pilgrims’ datasets to select only the pilgrims’ prominent frames for further processing. lightweight MobileNetV2 is transferred for learning violent activity recognition utilizing spatial and temporal features. For performance evaluation, we have conducted experiments on two publicly avail able datasets: Hockey Fight and Surveillance Fight Dataset.

29. Methodology

30. Dataset

31. Results

32. Hajj pilgrimage abnormal crowd movement monitoring using optical flow and FCNN (journal of Big Data 2023)

33. Introduction The identification of anomalies in scenes, a trained and supervised FCNN is turned into an FCNN using FCNNs and temporal data. By minimizing computational complexity, incorrect movement detection is utilized to achieve high performance in terms of speed and precision

34. Contributions This is the first time, to our knowledge, that an FCNN and optical flow have been used to discover abnormalities. We provide an unique FCNN architecture for rapidly identifying erroneous movement abnormalities. The proposed technique exceeds state-of-the-art approaches in terms of performance, it outperforms them in terms of time since most applications are real-time. On a typical GPU, we reached a processing speed of 370 frames per second, more than five times faster than the previous quickest approach.

35. Methodology

36. Dataset Hajj‑Crowd‑2021 anomaly dataset The dataset comprises 100 videos that are considered normal and 100 videos that are considered abnormal. Each video, whether normal or abnormal, has 60,000 frames. The whole scene was shot in high definition at a frame rate of 25Hz frames per second.

37. UCSD ped2 This dataset includes 12 video tests and 16 video tests for training. All test frames’ ground truth may be accessed to evaluate the localization. 2,384 abnormal and 2,566 normal frames make up the 2,384 abnormal frames.

38. Subway T his dataset contains two sequences gathered during the entrance (1 h and 36 min, 144,249 frames) and exit of a subway station (43 min, 64,900 frames). The majority of individuals that enter and exit the station behave nicely. Individuals moving in the wrong direction (i.e., leaving from or entering the entry) or escaping payment are classified as atypical instances.

39. Abnormal detection process (Labeling) Abnormality happens when the crowd movement does not follow the normal Tawaf movement. From a 10 s video, we extract the image frames. Select every frame and check the abnormal motion vector for each image. If the image has abnormal motion vectors of more than around 20% across 10 s, we label it abnormal. We have checked manually if the labelling has not been corrected.