Background Suppression Network for Weakly-Supervised Temporal Action Localization

Weakly-supervised temporal action localization is a very challenging problem because frame-wise labels are not given in the training stage while the only hint is video-level labels: whether each video contains action frames of interest. Previous methods aggregate frame-level class scores to produce video-level prediction and learn from video-level action labels. This formulation does not fully model the problem in that background frames are forced to be misclassified as action classes to predict video-level labels accurately. In this paper, we design Background Suppression Network (BaS-Net) which introduces an auxiliary class for background and has a two-branch weight-sharing architecture with an asymmetrical training strategy. This enables BaS-Net to suppress activations from background frames to improve localization performance. Extensive experiments demonstrate the effectiveness of BaS-Net and its superiority over the state-of-the-art methods on the most popular benchmarks – THUMOS'14 and ActivityNet. Our code and the trained model are available at https://github.com/Pilhyeon/BaSNet-pytorch.

Morphoagronomic characterization of sweet potato accessions from the germplasm bank of Embrapa hortaliças

Studies on the determination of genetic divergence among genotypes are important tools in breeding programs, contributing to the identification of parents with considerable productive potential. However, little is known about the combinatorial capacity of sweet potato (Ipomoea batatas) accessions and its adaptation to the different regions of Brazil. The objective of this study was to evaluate the morpho-agronomic traits from 102 sweet potato accessions from the Germplasm Bank of Embrapa Hortaliças. The experiment was laid out as an augmented block design comprised of 102 treatments. Nineteen above ground traits were measured using descriptors for the respective parts. Estimated values of broad sense heritability were high for the traits mean branch length (95.75%), immature leaf color (85.06%), and predominant branch color (90.57%). Coefficients of environmental variation were below 30.00% for all variables, except for branch weight (51.62%). The 102 clones analyzed presented broad genetic variability for the different traits evaluated, especially for branch weight, and branch length, and mature leaf color.

Statuses and double branch weights of quadrangular outerplanar graphs

In this paper we study some distance properties of outerplanar graphs with the Hamiltonian cycle whose all bounded faces are cycles isomorphic to the cycle C₄. We call this family of graphs quadrangular outerplanar graphs. We give the lower and upper bound on the double branch weight and the status for this graphs. At the end of this paper we show some relations between median and double centroid in quadrangular outerplanar graphs.

Statuses and double branch weights of quadrangular outerplanar graphs

In this paper we study some distance properties of outerplanar graphs with the Hamiltonian cycle whose all bounded faces are cycles isomorphic to the cycle C4. We call this family of graphs quadrangular outerplanar graphs. We give the lower and upper bound on the double branch weight and the status for this graphs. At the end of this paper we show some relations between median and double centroid in quadrangular outerplanar graphs

The Viterbi Decoding Scheme for FPGA

This paper analyzes the principle of Viterbi algorithm which can be used in the norm of the mobile communication system. Then a new Viterbi decoding scheme of (2, 1, 7) convolutional code is presented for FPGA implementation. To take advantage of the FPGA, a new branch weight algorithm and uniform state weight memories is used. At last, a new decoding circuit which can work on 35MHz and can achieve 120 kbs in decoding speed was designed. To use the design of survival path exchange register module, it can decrease the power consumption and the RAM size.

Spurred anoda (Anoda cristata) interference in wide row and ultra narrow row cotton

A field experiment was conducted in 2000, 2001, and 2002 at Stoneville, MS, to determine the effect of spurred anoda interference on yield loss of two cotton cultivars, ‘Delta Pine 5415’ and ‘Pima S-6’, grown under wide (1 m) (WR) and ultra narrow (0.25 m) row (UNR) spacings. The relationship between spurred anoda density and dry weight per plot was linear each year. At a spurred anoda density of 8 m−2, spurred anoda dry weight per plot was 507, 322, and 777 g m−2in 2000, 2001, and 2002, respectively. However, spurred anoda did not interfere with seed cotton yield in 2001, which was probably attributable to the low branch development in that year. Yield losses exceeded 55% at a spurred anoda density of 8 m−2compared with controls in both WR and UNR. The effect of spurred anoda density on boll numbers was nearly identical in 2000 and 2002, regardless of cotton cultivar and row spacing. Boll weights decreased in response to spurred anoda interference. Spurred anoda interference resulted in a decrease in cotton branch dry weight in WR but not in UNR. The yield decrease as a result of spurred anoda interference in WR was due to reduction in boll retention or fruiting sites (predicated on a decrease in branch weight). However, in UNR, the yield decrease was due to plant mortality; the plant density of both cotton cultivars decreased by one plant for each additional spurred anoda, but the yield per plant for surviving plants remained constant. Neither WR nor UNR cotton had significant advantage in response to spurred anoda interference. The decreased boll weight observed in UNR, and the failure to increase boll numbers m−2to compensate for decreased boll weight in UNR compared with WR, may limit its appeal to cotton producers.

How does density and nutrient stress affect allometry and fruit production in the heterocarpic species Atriplex sagittata (Chenopodiaceae)?

Plants of Atriplex sagittata Borkh. were grown under different density and (or) nutrient conditions, and the effect of stress on plant allometry and variation in particular biometrical parameters was studied. Increasing density significantly reduced total weight, basal diameter, stem length, and number of branches. Nutrient deficiency had a negative effect on the total weight and basal diameter only, and the density × nutrients interactions were all nonsignificant. Basal diameter was the best predictor of the total weight, explaining 93.6% of the variation in the whole data set. The relationship between total plant weight and basal diameter was unaffected by growth conditions, indicating that competition determined the position of a plant along the line describing the allometric relationship between both variables but did not seem to alter the relationship. However, the allometric relationships between total weight and stem height and between total weight and number of branches found at low density were significantly different from those at high density, indicating a faster increase in total weight at low densities. Branching, and consequently plant form, were affected by competition. Branch length consistently increased with total plant weight in plants grown under high density, but for low density plants, such an increase was recorded only in branches of the middle section of the stem. In each fruit type, the total weight of fruits produced was more closely predicted by branch weight than by branch length. Allometric relationships between fruit production and branch weight differed among particular fruit types, with a faster increase in the production of type A than in that of types B or C. Competition had less effect on the mean size of type B fruits compared with that of types A or C. Fruits of types C and A had higher variation in mean weight, than type B fruit. Variation in total weight of fruits per plant was of a higher order of magnitude than that in the mean fruit weight. The results show that plants of A. sagittata grown under contrasting conditions change both their allometry and the number of particular fruit types produced. Nevertheless, the question of whether the shift in the fruit ratio is a function of allometric constraints or the result of a basic shift in allometry needs further study.Key words: allometry, Atriplex sagittata, Chenopodiaceae, density stress, nutrient stress, heterocarpy.