Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Indrit Enesi has a total of 16 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    The Role of Camera Intrinsic Initialization in End-to-End SfM/MVS Reconstruction: A Comparative Evaluation of Metadata, Fallback, and Learned Calibration

    by Anduel Kuqi, Ambra Korra, Indrit Enesi

    For structure-from-motion (SfM), an important step is provided through camera intrinsics, especially when camera metadata are not available. Although subsequent SfM optimization and bundle adjustment (BA) refine the camera parameters using multi-view observations, it remains unclear whether differences in the initial intrinsics have a lasting effect on the optimization process and, ultimately, on the quality of the reconstructed scene. This study investigates the role of camera intrinsic initialization in an end-to-end AliceVision/Meshroom SfM and multi-view stereo (MVS) pipeline, with particular focus on whether similar optimized intrinsics necessarily lead to similar final reconstructions. Here, three initialization strategies are evaluated within an AliceVision/Meshroom pipeline: camera intrinsics obtained from image metadata, AliceVision’s 45° field-of-view fallback, and single-image calibration using a pretrained GeoCalib model. GeoCalib is integrated only at the camera-initialization stage, where its predicted intrinsics are written into the AliceVision SfMData structure. All subsequent stages of the reconstruction pipeline remain unchanged, including feature extraction, image matching, feature matching, SfM, bundle adjustment, dense reconstruction, meshing, and texturing. Experiments are conducted on the 41-image Monstree smartphone collection and the 27-image ETH3D Observatory sequence. ETH3D additionally provides reference camera calibration and sparse reconstruction data, allowing the effect of initialization to be measured against ground truth. Here, the evaluation is extended to selected six-view, eight-view, and twelve-view Observatory subsets, the 26-view Boulders scene, and paired dense reconstructions of the 23-view ETH3D Terrace training scene. The experiments show that the three methods can produce considerably different initial camera parameters while still converging to similar optimized focal length. For the evaluated datasets, this suggests that the final intrinsic parameters are strongly influenced by multi-view optimization rather than by the initialization alone. The experiments also show that more accurate single-image calibration does not necessarily result in a correspondingly different or improved optimized focal length. However, similarity in the final intrinsic parameters does not, by itself, demonstrate that the resulting dense reconstructions are equivalent. The six-view and Boulders comparisons show that the convergence observed on the full Observatory sequence does not extend to every tested configuration. Terrace achieves F1-scores of 0.8273 for fallback and 0.8261 for learned initialization at 2 cm against public laser scans. These findings highlight the need to distinguish between the quality of the initial camera calibration, the recovery of camera parameters during SfM optimization, and the quality of the final dense reconstruction. These findings remain limited to the selected subsets and tested scenes.

  • Open Access

    Artificial Intelligence for Intracranial Aneurysm Detection, Rupture-Risk Prediction, and Prognosis Evaluation: Recent Advances and Clinical Integration

    by Olti Qirici, Eugen Enesi, Kleona Binjaku, Anduel Kuqi, Indrit Enesi, Ana Ktona, Elinda Kajo Meçe, Arjan Durresi

    Intracranial aneurysm rupture is a major cause of hemorrhagic stroke and a significant contributor to human mortality. With the emergence of artificial intelligence (AI), particularly machine learning (ML) and deep learning (DL), a series of new approaches have been introduced to improve aneurysm detection, rupture assessment, risk prediction, and prognosis evaluation. This narrative review synthesizes the findings from 21 peer-reviewed studies identified through searches in PubMed, Scopus, Web of Science, and IEEE Xplore, covering the period from June to September 2025. The reviewed studies examined the use of AI for aneurysm detection, rupture assessment and risk prediction, and prognosis evaluation based on clinical, morphological, radiomic, and hemodynamic data. The reviewed studies applied different AI approaches, including conventional ML algorithms, convolutional neural networks, transformer-based models, radiomics, and multimodal models. Several studies reported encouraging performance in aneurysm detection, segmentation, rupture assessment, and prognosis evaluation, while some approaches showed better performance compared with traditional clinical scoring systems. However, important limitations remain, including the frequent use of small or single-center datasets, limited external validation, and heterogeneity in imaging protocols and feature-extraction methods. AI is showing strong potential in supporting the management of intracranial aneurysms. However, further research is needed, particularly through multicenter and external validation, improved model explainability and robustness, and better integration of these systems into clinical workflows.

  • Open Access

    Hyperparameter Tuning of Inception CNNs Using Genetic Algorithms for Automatic Defect Detection

    by Ambra Korra, Anduel Kuqi, Indrit Enesi

    Automated defect detection in industrial casting processes is important for improving product quality while reducing the cost of manual inspection. In this work, two deep convolutional neural network (CNN) architectures, InceptionV3 and InceptionResNetV2, are evaluated for the binary classification of defects in submersible pump impellers. A genetic algorithm (GA) is used to optimize key hyperparameters, including dropout rate, learning rate, and dense layer configuration, while model complexity is assessed through Pareto-based analysis. Single-run optimization results show that InceptionV3 achieves high classification accuracy (99.0%) with lower model complexity than InceptionResNetV2 (98.75%). Repeated experiments using different random seeds demonstrate relatively stable performance across runs, with InceptionV3 achieving an accuracy of 0.9913 ± 0.003 and InceptionResNetV2 achieving 0.9860 ± 0.0076. Additional experiments were conducted using random-search baselines and classification-head ablation studies (Flatten vs. Global Average Pooling). These experiments showed that optimization strategy and architectural design choices influence both predictive performance and computational complexity. The environmental impact of the training process is evaluated using CodeCarbon, with energy consumption ranging from 0.083 to 0.098 kWh and carbon emissions ranging from 2.008 to 2.401 g CO2eq for InceptionV3 and InceptionResNetV2, respectively. Overall, the results suggest that the most effective configuration depends on the evaluated architecture and experimental setting, highlighting the importance of balancing accuracy, model complexity, and computational efficiency in industrial defect detection systems.

  • Open Access

    Performance Analysis for 3D Reconstruction Objects in Meshroom and Agisoft—A Comparative Study

    by Indrit Enesi, Anduel Kuqi

    3D reconstruction of objects is with ​​interest nowadays, mainly in production industry. The main challenge of them is accuracy and processing time, especially for small and detailed objects. The field of photogrammetry realizes the 3D reconstruction of objects through 2D photos. Different software, free or non-free exists, providing different quality and performance. Accurate 3D reconstruction is important in cloning objects, especially in the industry of spare parts or in the production of prostheses in medicine, etc. Determining accurately the sizes of the object, especially those with complex geometric shapes is very important in the 3D printing process. The purpose of this paper is the analysis of the performance of 3D reconstruction in terms of accuracy for objects of different sizes regarding the number of its photos and the time evaluation of this process. The 3D reconstruction will be performed by free software Meshroom, measurement will be done in MeshLab and non-free software Agisoft. Experimental results show that quality and performance of 3D reconstruction depends on the number of photos of the object, concluding in finding the optimal balance between these parameters. By comparing obtained results from MeshRoom and Meshlag versus Agisoft, it is claimed that Agisoft performs better than MeshRoom. It offers more optimization techniques, reduces processing time, more visual quality in the reconstructed 3D object as well as more accuracy in measurement.


  • Open Access

    Analyzing the Measurement Accuracy of 3D Reconstruction for Small Objects using Different Software; A Case Study

    by Indrit Enesi, Anduel Kuqi

    Analyzing the Measurement Accuracy of 3D Reconstruction for Small Objects using Different Software; A Case Study by Indrit Enesi* , Anduel Kuqi Polytechnic University of Tirana, Electronic and Telecommunication Department, Faculty of Information Technology, Tirana, 1000, Albania * Author to whom correspondence should be addressed. Journal of Engineering Research and Sciences, Volume 1, Issue 12,


  • Open Access

    Implementing Fusion Technique Using Biorthogonal Dwt to Increase the Number of Minutiae in Fingerprint Images

    by Indrit Enesi, Miranda Harizaj, Betim Çiço

    Biometric devices identify persons based on the minutiae extracted from fingerprint images. Image quality is very important in this process. Usually, fingerprint images have low quality and in many cases they are obtained in various positions. The paper focuses on increasing minutiae detected number by fusing two fingerprint images obtained in various positions. Biorthogonal wavelets have advantages compared to orthogonal wavelets. Fusion is performed in wavelet domain by implementing biorthogonal wavelet. Terminations and bifurcations are extracted from the original and fused images using licensed software Papillon 9.02 and manually extraction by an expert. Biorthogonal Wavelet transform is implemented in the image fusion process, yielding in the increased number of the minutiae compared to the original one. Different biorthogonal wavelets are experimented and various results are obtained. Finding the appropriate wavelet is important in the fusion process since it has a direct impact in the number of minutiae extracted. Based on the number of minutiae and MSE results, the appropriate wavelet to be used in the fusion process is defined.


  • Open Access

    Increasing quality in fingerprint images by Implementing wavelet transform based fusion technique

    by Indrit Enesi, Betim Çiço, Miranda Harizaj

    Fingerprint images are practically most widely used in biometric systems for person identification. Their quality is very important in extracting true minutiae. In this paper we treat the method of increasing the true minutiae detected number based on the fusion of two fingerprint images obtained in different positions. Wavelet transform based fusion method are implemented using different wavelets. From the results it yields that the fusion based on DWT transform increases the accuracy of minutiae detected.


  • Open Access

    Analysing Clustering Algorithms Performance in CRM Systems

    by Indrit Enesi, Ledion Liço, Aleksander Biberaj, Desar Shahu