Peer-Reviewed Journal & Conference Papers
-
Deep-BrownConrady: Prediction of Camera Calibration and Distortion Parameters Using Deep Learning and Synthetic Data
F. M. Chaudhry, J. Ralli, J. Leudet, F. Sohrab, F. Pakdaman, P. Corbani and M. Gabbouj — IEEE Transactions on Automation Science and Engineering, 2025
Read Abstract
This research addresses the challenge of camera calibration and distortion parameter prediction from a single image using deep learning models. The main contributions of this work are: (1) demonstrating that a deep learning model, trained on a mix of real and synthetic images, can accurately predict camera and lens parameters from a single image, and (2) developing a comprehensive synthetic dataset using the AILiveSim simulation platform. This dataset includes variations in focal length and lens distortion parameters, providing a robust foundation for model training and testing. The training process predominantly relied on these synthetic images, complemented by a small subset of real images, to explore how well models trained on synthetic data can perform calibration tasks on real-world images. Traditional calibration methods require multiple images of a calibration object from various orientations, which is often not feasible due to the lack of such images in publicly available datasets. A deep learning network based on the ResNet architecture was trained on this synthetic dataset to predict camera calibration parameters following the Brown-Conrady lens model. The ResNet architecture, adapted for regression tasks, is capable of predicting continuous values essential for accurate camera calibration in applications such as autonomous driving, robotics, and augmented reality. Note to Practitioners—This paper introduces a deep learning approach to predict camera calibration and distortion parameters directly from a single image, addressing the limitations of traditional methods that require structured calibration objects and multiple images. Using synthetic datasets generated with a simulation platform, the model predicts essential parameters such as field of view, principal points, and Brown-Conrady distortion coefficients. A key innovation is incorporating image size into the learning process, enabling the model to generalize well to real-world scenarios. This approach simplifies the calibration process, making it suitable for dynamic and unstructured environments, such as autonomous driving and robotics, where traditional calibration methods are not feasible. As an important result, the proposed method enables the use of synthetic data to overcome data scarcity in real-world applications by adapting to the camera parameters of the underlying physical system. By leveraging synthetic data and deep learning, the method offers a modern, flexible alternative that practitioners can adopt to enhance camera calibration workflows in real-world applications. Extensive experiments are reported that showcase a successful solution based on the famous Brown-Conrady camera lens model and are validated on a real-world dataset. We believe a similar methodology can be used and extended as future work to enable camera parameter estimation and the use of synthetic data for other camera models. -
From Sensors to Spikes: Evolving Receptive Fields to Enhance Sensorimotor Information in a Robot-arm
N. Luque, J. A. Garrido, J. Ralli, J. J. Laredo and E. Ros — International Journal of Neural Systems, 2012
Read Abstract
In biological systems, instead of actual encoders at different joints, proprioception signals are acquired through distributed receptive fields. In robotics, a single and accurate sensor output per link (encoder) is commonly used to track the position and the velocity. This paper analyzes how evolved receptive fields (optimized towards information transmission) can efficiently generate a sensorimotor representation that facilitates its discrimination from other "sensorimotor states." -
Low-cost Sensor to Detect Overtaking Based on Optical Flow
P. Guzman, J. Díaz, J. Ralli, R. Agís and E. Ros — Machine Vision and Applications, 2011
Read Abstract
The automotive industry invests substantial amounts of money in driver-security and driver-assistance systems. We propose an overtaking detection system based on visual motion cues that combines feature extraction, optical flow, solid-objects segmentation and geometry filtering, working with a low-cost compact architecture based on one focal plane and an on-chip embedded processor. -
Spatial and Temporal Constraints in Variational Correspondence Methods
J. Ralli, J. Díaz and E. Ros — Machine Vision and Applications, 2011
Read Abstract
In this paper we describe a novel use for a well known temporal constraint term in the framework of variational correspondence methods. The new use, that we call spatial constraining, allows bounding of the solution based on what is known of the solution beforehand. This spatial constraint term enables fusion of information between high- and low-level vision systems. -
Disparity Disambiguation by Fusion of Signal- and Symbolic-level Information
J. Ralli, J. Díaz, S. Kalkan, N. Krüger and E. Ros — Machine Vision and Applications, 2010
Read Abstract
We describe a method for resolving ambiguities in low-level disparity calculations in a stereo-vision scheme by using a recurrent mechanism that we call signal-symbol loop. Due to the local nature of low-level processing it is not always possible to estimate the correct disparity values. Symbolic abstraction of the signal produces robust, high confidence, multimodal image features which can be used to interpret the scene more accurately. -
A Method for Sparse Disparity Densification Using Voting Mask Propagation
J. Ralli, J. Díaz and E. Ros — Journal of Visual Communication and Image Representation, 2009
Read Abstract
We describe a novel method for propagating disparity values using directional masks and a voting scheme. The driving force of the propagation direction is image gradient, making the process anisotropic, whilst ambiguities between propagated values are resolved using a voting scheme. -
Increasing Efficiency in Disparity Calculation
J. Ralli, F. J. Pelayo — Advances in Brain, Vision, and Artificial Intelligence, 2007
Read Abstract
In this paper a trade-off between the computation effort and the accuracy of the resulting disparity map, obtained using interpolation over spatial domain, is presented. The accuracy of the obtained disparity map is presented as the mean squared error calculated over the known disparity ground truth of test images, while efficiency increase is presented as the reduction in computational time. We show that interpolation can be employed for increasing the disparity map computation efficiency and for deducing a disparity value where not present. In many real-time applications, an approximation of the true disparity map is accurate enough, and such systems can benefit considerably from the computation efficiency increment.
Conference Abstracts & Poster Presentations
-
Are Callipers Obsolete? A Novel 3D Scanning Technology to Measure Subcutaneous Tumor Volume
Z. Wilson, J. Delgado, M. Davies, R. Whiteley, J. Hare, A. Rahi, S. Marshall, A. Smith, S. Atkinson, J. Ralli, A. Zabair, A. Zabair, J. Kendrew — American Association for Laboratory Animal Science, 2017
Read Abstract
Most preclinical oncology studies (xenograft, PDX, GEMMS) involve monitoring tumor growth rates, measuring them with calipers, and calculating the volume. Volume is calculated from the width and the length to estimate a 3D volume and is directly used to assess treatment efficacy. Although this technique is useful, it is unable to accurately assess nonuniformly shaped or small tumors and introduces a systematic bias by assuming that tumors present with spheroid shape. We describe the development and validation of a 3D scanner as an alternative method to calipers to monitor tumor progression in rodents. The resulting 3D scanner has the potential to deliver significant 3Rs benefits, identified as reduction of animals via improved data accuracy allowing reduction in group sizes or the ability to include irregularly shaped tumors to test. In addition, the scanner system described will make it possible to record tumor measurements in a rapid, minimally invasive, morphology-independent, and human bias-free way, removing interoperator variability. We describe the development and validation of the scanner system within our laboratories, evaluating the finalprototype hardware. Using the 3D scanner alongside tumor calipers to monitor tumor growth of oncology tumorstudies, we demonstrate that we can measure tumor size parameters (length, width, and volume), in multiple mouse strains and across a range of tumor models, with accuracy and precision comparable to tumor measures generated from calipers. If successful the introduction of this system to replace tumor calipers could have a large impact for groups running oncology in vivo tumor studies. It operates with different sized aperture holes to accommodate different rodents and tumors. The process of taking a measurement includes taking the rodent by the scruff (in the same way as with calipers), placing the rodents tumor in the center of the scanning window, and then capturing the 3D data by pushing the scan button on the device or in the software. The software then segments the tumor for size and shape data which is then exported for analysis over time. Mouse strains tested internally include immunocompromised athymic nude, SCID and NSG mice, and immunocompetent Balb/C and C57BL6 mice. -
Optical Flow for Motion Estimation and Tracking of Subcellular, Cellular and Supracellular Dynamics
M. Cerda, J. Jara, A. Córdova, J. Toledo, E. Pulgar, C-G Lemus, O. Ramírez, J. Ralli, Miguel Concha, Steffen Härtel — Chilean Society for Cell Biology XXVI Annual Meeting, 2012
Read Abstract
Cell migration, formation of cellular protrusions (e.g. blebs, filopodia), and structural reorganization are important phenomena in cell biology. Precise quantifications of movement and deformation are required to understand the underlying biomechanical processes. We apply optical flow techniques to estimate these complex dynamic behaviors across different cellular scales.
PhD Thesis
-
Fusion and Regularisation of Image Information in Variational Correspondence Methods
J. Ralli — University of Granada, Spain, 2011
Read Abstract
In this work we study, and improve, applicability of variational correspon- dence methods, used for calculating dense optical-flow and disparity fields, to real scenes with realistic illumination conditions. It is well known that under realistic illumination conditions and image noise, proper image rep- resentation is crucial in order to generate both correct and temporally co- herent optical-flow and disparity fields. We have studied 34 different image representations and ranked these with respect to accuracy, robustness and a combination of accuracy plus robustness. In the case of well known test im- ages with optimal (or quasi optimal) illumination conditions, effects of im- age representation are not that important. On the other hand, with scenes from real applications, such as robotic grasping or vision in automotive in- dustry, influence of image representation is crucial. Also we have extended the basic models to include both spatial- and temporal-constraints. In the case of optical-flow, for example, temporal constraint reduces ‘flickering’ of estimations. By flickering we mean temporal changes in the displacement fields due to lack of or ambiguity of spatial features in the images. We show that by using spatial constraints in the disparity estimation, considerable improvements are possible. These constraints are due to (a) what we know of the solution before hand (e.g. roads are relatively flat surfaces, sky is far away) or (b) what we can deduce from the scene itself. Effectively, we show how these constraints can be obtained and refined in a hypothesis-forming- validation loop (HFVL). In the example that we give of a HFVL loop, we segment an initial disparity map and form constraint(s) based on the segments and feed back these into the disparity calculation.Not only do the disparity estimations improve, but we also segment the depth map into meaningful surfaces Apart from introducing the principal results obtained, we also explain in detail how the models that we have used can be solved. Therefore, this work (along with the related Matlab/MEX code available at http://http: //www.jarnoralli.fi/joomla/code) will hopefully help other scientists understand and further improve the the variational methods for calculating stereo-disparity and optical-flow.