Parametric Image Alignment Using Enhanced Correlation Coefficient Maximization

Signal Processing and Communications Lab, Department of Computer Engineering and Informatics, University of Patras, Rio-Patras, Greece.
IEEE Transactions on Pattern Analysis and Machine Intelligence (Impact Factor: 5.78). 11/2008; 30(10):1858-65. DOI: 10.1109/TPAMI.2008.113
Source: PubMed


In this work we propose the use of a modified version of the correlation coefficient as a performance criterion for the image alignment problem. The proposed modification has the desirable characteristic of being invariant with respect to photometric distortions. Since the resulting similarity measure is a nonlinear function of the warp parameters, we develop two iterative schemes for its maximization, one based on the forward additive approach and the second on the inverse compositional method. As it is customary in iterative optimization, in each iteration, the nonlinear objective function is approximated by an alternative expression for which the corresponding optimization is simple. In our case we propose an efficient approximation that leads to a closed-form solution (per iteration) which is of low computational complexity, the latter property being particularly strong in our inverse version. The proposed schemes are tested against the Forward Additive Lucas-Kanade and the Simultaneous Inverse Compositional (SIC) algorithm through simulations. Under noisy conditions and photometric distortions, our forward version achieves more accurate alignments and exhibits faster convergence whereas our inverse version has similar performance as the SIC algorithm but at a lower computational complexity.

Download full-text


Available from: Emmanouil Psarakis, Oct 03, 2015
1 Follower
46 Reads
  • Source
    • "Image analysis was performed off line using bespoke MATLAB based software (MATLAB 2013, MathWorks, Natick, Massachusetts, U.S.A). Registration of the ASL images was performed using an enhanced correlation coefficient maximisation algorithm [14]. For each pair of selective/non-selective inversion images, the non-selective inversion image was subtracted from the selective inversion image. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Measurement of renal perfusion is a crucial part of measuring kidney function. Arterial spin labelling magnetic resonance imaging (ASL MRI) is a non-invasive method of measuring renal perfusion using magnetised blood as endogenous contrast. We studied the reproducibility of ASL MRI in normal volunteers. ASL MRI was performed in healthy volunteers on 2 occasions using a 3.0 Tesla MRI scanner with flow-sensitive alternating inversion recovery (FAIR) perfusion preparation with a steady state free precession (True-FISP) pulse sequence. Kidney volume was measured from the scanned images. Routine serum and urine biochemistry were measured prior to MRI scanning. 12 volunteers were recruited yielding 24 kidneys, with a mean participant age of 44.1 +/- 14.6 years, blood pressure of 136/82 mmHg and chronic kidney disease epidemiology formula estimated glomerular filtration rate (CKD EPI eGFR) of 98.3 +/- 15.1 ml/min/1.73 m2. Mean kidney volumes measured using the ellipsoid formula and voxel count method were 123.5 +/- 25.5 cm3, and 156.7 +/- 28.9 cm3 respectively. Mean kidney perfusion was 229 +/- 41 ml/min/100 g and mean cortical perfusion was 327 +/- 63 ml/min/100 g, with no significant differences between ASL MRIs. Mean absolute kidney perfusion calculated from kidney volume measured during the scan was 373 +/- 71 ml/min. Bland Altman plots were constructed of the cortical and whole kidney perfusion measurements made at ASL MRIs 1 and 2. These showed good agreement between measurements, with a random distribution of means plotted against differences observed. The intra class correlation for cortical perfusion was 0.85, whilst the within subject coefficient of variance was 9.2%. The intra class correlation for whole kidney perfusion was 0.86, whilst the within subject coefficient of variance was 7.1%. ASL MRI at 3.0 Tesla provides a repeatable method of measuring renal perfusion in healthy subjects without the need for administration of exogenous compounds. We have established normal values for renal perfusion using ASL MRI in a cohort of healthy volunteers.
    BMC Nephrology 01/2014; 15(1):23. DOI:10.1186/1471-2369-15-23 · 1.69 Impact Factor
  • Source
    • "An interpolation scheme on the w i all provides pixel-wise weights and the latter, after their normalization, apply to both images. Other than the self-weighting, the steps of the algorithm exactly follow the original version [13]. It is important to note that there is no compositional weighting, but the images, i.e., I n and warped I m , are re-weighted at each iteration. "
    [Show abstract] [Hide abstract]
    ABSTRACT: This paper addresses the background estimation problem for videos captured by moving cameras, referred to as video grounding. It essentially aims at reconstructing a video, as if it would be without foreground objects, e.g. cars or people. What differentiates video grounding from known background estimation methods is that the camera follows unconstrained motion so that background undergoes ongoing changes. We build on video matching aspects since more videos contribute to the reconstruction. Without loss of generality, we investigate a challenging case where videos are recorded by in-vehicle cameras that follow the same road. Other than video synchronization and spatiotemporal alignment, we focus on the background reconstruction by exploiting inter- and intra-sequence similarities. In this context, we propose a Markov random field formulation that integrates the temporal coherence of videos while it exploits the decisions of a support vector machine classifier about the background ness of regions in video frames. Experiments with real sequences recorded by moving vehicles verify the potential of the video grounding algorithm against state-of-art baselines.
    2013 IEEE International Conference on Computer Vision Workshops (ICCVW); 12/2013
  • Source
    • "To summarize, once a temporal registration has been initialized by means of the efficient retrieval step, the ECC algorithm converges to that subframe that maximizes the value of the enhanced correlation coefficient [3]. In other words, the highest number of feature matches provides a rough synchronization, and the maximum similarity defines the final corresponding reference frame. "
    [Show abstract] [Hide abstract]
    ABSTRACT: This paper addresses the problem of video alignment. We present efficient approaches that allow for spatiotemporal alignment of two sequences. Unlike most related works, we consider independently moving cameras that capture a 3D scene at different times. The novelty of the proposed method lies in the adaptation and extension of an efficient information retrieval framework that casts the sequences as an image database and a set of query frames, respectively. The efficient retrieval builds on the recently proposed quad descriptor. In this context, we define the 3D Vote Space (VS) by aggregating votes through a multiquerying (multiscale) scheme and we present two solutions based on VS entries; a causal solution that permits online synchronization and a global solution through multiscale dynamic programming. In addition, we extend the recently introduced ECC image-alignment algorithm to the temporal dimension that allows for spatial registration and synchronization refinement with subframe accuracy. We investigate full search and quantization methods for short descriptors and we compare the proposed schemes with the state of the art. Experiments with real videos by moving or static cameras demonstrate the efficiency of the proposed method and verify its effectiveness with respect to spatiotemporal alignment accuracy.
    IEEE Transactions on Pattern Analysis and Machine Intelligence 10/2013; 35(10):2371-86. DOI:10.1109/TPAMI.2013.56 · 5.78 Impact Factor
Show more