We address the problem of estimating the position and motion of a human arm in 3 0 without any constraints on its behavior and without the use of special markers. We model the arm as two truncated right-circular cones connected with spherical joints. We propose to use a recursive estimator for arm position, and to provide the estimator with error signals obtained by comparing the projected estimated arm position with that of the actual arm in the image. The system is demonstmted and tested on a real image sequence.

& Key message Industrial computed tomography scanning of logs provides detailed information on timber quality prior to sawing. A sawing simulation-considering log rotation angle and knot size accuracy-revealed an average value increase of up to 20% for the best angle compared to the conventional horns-up position. & Context Computed tomography (CT) scanning has the potential to improve the value of products sawn from logs and meets the increasing demands of the wood industry for detailed information on log quality prior to processing. & Aims In a validation step, automated measurements of knot cluster variable DAB (DIN 4074-1:2012-06) using CT were compared with manual measurements. In a second optimization step, the hypothesis that the value of the sawn products is increased by sawing at the best rotation angle as opposed to the horns-up position was tested. & Methods A sample of 36 Douglas-fir logs were scanned in an industrial CT scanner, and sawn into boards. Knots on the boards were manually measured, and compared with the corresponding knots on virtual boards created from the CT data. The error of the DAB was measured by comparing CT data to manual measurements. An optimized sawing simulation was performed, using the measured DAB error to account for CT measurement errors, as well as a rotational error to account for errors in the log turning equipment. Using the results of the sawing simulation, Monte Carlo simulations were performed to show the potential and benefit of an industrial CT scanner. & Results The three largest DABs measured by the CT showed good correlation to the measurements on the manual boards. The simulation revealed an average increase of value from 4 to 20% compared to the conventional horns-up position depending on the relative price differences between the strength grades. & Conclusion By using a CT scanner to optimize sawing, sawmill owners can process logs in a better way to produce final products with increased added value.

Within modern Deep Learning setups, data augmentation is the weapon of choice when dealing with narrow datasets or with a poor range of different samples. However, the benefits of data augmentation are abysmal when applied to a dataset which is inherently unable to cover all the categories to be classified with a significant number of samples. To deal with such desperate scenarios, we propose a possible last resort: Cross-Dataset Data Augmentation. That is, the creation of new samples by morphing observations from a different source into credible specimens for the training dataset. Of course specific and strict conditions must be satisfied for this trick to work. In this paper we propose a general set of strategies and rules for Cross-Dataset Data Augmentation and we demonstrate its feasibility over a concrete case study. Even without defining any new formal approach, we think that the preliminary results of our paper are worth to produce a broader discussion on this topic.