scholarly journals Automatic Detection and Segmentation for Group-Housed Pigs Based on PigMS R-CNN

Sensors ◽  
2021 ◽  
Vol 21 (9) ◽  
pp. 3251
Author(s):  
Shuqin Tu ◽  
Weijun Yuan ◽  
Yun Liang ◽  
Fan Wang ◽  
Hua Wan
Keyword(s):  
Body Condition Score ◽  
Live Weight ◽  
Segmentation Method ◽  
Feature Maps ◽  
Residual Network ◽  
Candidate Network ◽  
Welfare Evaluation ◽  
Condition Score ◽  
Feature Pyramid ◽  
Instance Segmentation

Instance segmentation is an accurate and reliable method to segment adhesive pigs’ images, and is critical for providing health and welfare information on individual pigs, such as body condition score, live weight, and activity behaviors in group-housed pig environments. In this paper, a PigMS R-CNN framework based on mask scoring R-CNN (MS R-CNN) is explored to segment adhesive pig areas in group-pig images, to separate the identification and location of group-housed pigs. The PigMS R-CNN consists of three processes. First, a residual network of 101-layers, combined with the feature pyramid network (FPN), is used as a feature extraction network to obtain feature maps for input images. Then, according to these feature maps, the region candidate network generates the regions of interest (RoIs). Finally, for each RoI, we can obtain the location, classification, and segmentation results of detected pigs through the regression and category, and mask three branches from the PigMS R-CNN head network. To avoid target pigs being missed and error detections in overlapping or stuck areas of group-housed pigs, the PigMS R-CNN framework uses soft non-maximum suppression (soft-NMS) by replacing the traditional NMS to conduct post-processing selected operation of pigs. The MS R-CNN framework with traditional NMS obtains results with an F1 of 0.9228. By setting the soft-NMS threshold to 0.7 on PigMS R-CNN, detection of the target pigs achieves an F1 of 0.9374. The work explores a new instance segmentation method for adhesive group-housed pig images, which provides valuable exploration for vision-based, real-time automatic pig monitoring and welfare evaluation.

scholarly journals 178 Use of silvopastures and woodlands to expand the grazing opportunity for small ruminants

2020 ◽  
Vol 98 (Supplement_2) ◽  
pp. 67-68
Author(s):  
Uma Karki
Keyword(s):  
Body Condition Score ◽  
Small Ruminants ◽  
Warm Season ◽  
Live Weight ◽  
Limited Resource ◽  
Ground Level ◽  
Vegetation Biomass ◽  
Meat Goats ◽  
Condition Score ◽  
And Behavior

Abstract Pasture-based small-ruminant farming is a popular enterprise in the Southeast, especially for small and limited-resource producers. Although woodlands occupy a majority of the land cover in the South and can be a promising resource for raising small ruminants, not much attention has been given on using such resources. A series of studies were conducted at the facilities of Tuskegee University with the overall objective of exploring the potential of using silvopastures and woodlands for raising small ruminants. Grazing studies were conducted from 2015 to 2019 using meat goats and hair sheep in silvopastures and woodland plots. Silvopastures were developed by thinning down the existing woodlands and planting suitable cool- and warm-season forages. Grazing facilities were installed in the study plots and animals were rotationally stocked each year during both cool- and warm-season grazing periods. In 2017, non-pine plants in woodlands were cut to three different heights (ground level and 0.9 m and 1.5 m from the ground level) or left uncut (control) to see the effects on understory vegetation biomass production and utilization by small ruminants. Data on vegetation biomass and quality, canopy height, animal performance and behavior, browsing height, and vegetation preference were collected and analyzed. Small ruminants were found to utilize all planted forages (silvopastures) well and most of the understory plants in woodlands. Vegetation biomass in woodlands increased in areas where non-pine plants were cut versus the control (36–106%; P < 0.0001). Mature animals maintained a desirable body condition score (≥2.6) and FAMACHA score (≤2.6) in woodlands. However, young, growing animals showed a poor live weight gain, especially during some portion of the study. Animals performed well when they were stocked in silvopastures. Both species showed a similar preference for most plant species available in woodlands. Results show a tremendous potential of using silvopastures and woodlands for expanding the grazing opportunity for small ruminants. However, the provision of some supplements would be necessary while stocking young animals in woodlands.

The effect of ammonia treatment on the intake of straw and hay when offered with rations of concentrates to ewes in late pregnancy

1985 ◽  
Vol 40 (1) ◽  
pp. 101-109 ◽  
Author(s):  
R. J. Orr ◽  
T. T. Treacher ◽  
V. C. Mason
Keyword(s):  
Organic Matter ◽  
Body Condition Score ◽  
Spring Barley ◽  
Live Weight ◽  
Late Pregnancy ◽  
Barley Straw ◽  
Ammonia Treatment ◽  
Condition Score ◽  
Forage Intake

ABSTRACTFinnish Landrace × Dorset Horn ewes were offered 300, 600 or 900 g fresh weight per day of concentrates and forage ad libitum from day 105 of pregnancy until lambing. Spring barley straw (S) or hay (H) was offered either untreated (U) or following treatment with anhydrous ammonia in an oven (T). Organic matter digestibilities (in vitro) were 0·42, 0·58, 0·42 and 0·60 and nitrogen contents were 7·2, 18·6, 12·0 and 25·0 g/kg dry matter for US, TS, UH and TH respectively. Forage intake did not differ between ewes carrying two or more foetuses but the small number of ewes carrying one foetus ate more straw (6·8 v. 4·5 g organic matter (OM) per kg live weight) than ewes carrying two or more foetuses. Ammonia treatment increased intake; the increase was larger on straw (4·6 v. 100 g OM per kg live weight) than on hay (9·0 v. 10·7 g OM per kg live weight). Replacement rates of forage by concentrates were -0·21, +0·06, -0·48 and +0·08 kg forage per kg concentrates for treatments US, TS, UH and TH respectively; only the value for treatment UH differed significantly from zero. On most treatments forage intake decreased as pregnancy progressed and the declines were greater when treated forages were offered. Concentrate level had a large effect on most aspects of ewe performance. Ewes offered treated forage gained slightly more weight in pregnancy (138 v. 104 g/day), had a slightly smaller decrease in body condition score (-0·54 v. -0·68) between day 105 and lambing but did not have greater lamb birth weights than ewes on untreated forage.

scholarly journals An Efficient Module for Instance Segmentation Based on Multi-Level Features and Attention Mechanisms

2021 ◽  
Vol 11 (3) ◽  
pp. 968
Author(s):  
Yingchun Sun ◽  
Wang Gao ◽  
Shuguo Pan ◽  
Tao Zhao ◽  
Yahui Peng
Keyword(s):  
Feature Extraction ◽  
Spatial Structure ◽  
Semantic Feature ◽  
Semantic Features ◽  
Segmentation Method ◽  
Spatial Dimensions ◽  
Feature Pyramid ◽  
Multi Level ◽  
High Level ◽  
Instance Segmentation

Recently, multi-level feature networks have been extensively used in instance segmentation. However, because not all features are beneficial to instance segmentation tasks, the performance of networks cannot be adequately improved by synthesizing multi-level convolutional features indiscriminately. In order to solve the problem, an attention-based feature pyramid module (AFPM) is proposed, which integrates the attention mechanism on the basis of a multi-level feature pyramid network to efficiently and pertinently extract the high-level semantic features and low-level spatial structure features; for instance, segmentation. Firstly, we adopt a convolutional block attention module (CBAM) into feature extraction, and sequentially generate attention maps which focus on instance-related features along the channel and spatial dimensions. Secondly, we build inter-dimensional dependencies through a convolutional triplet attention module (CTAM) in lateral attention connections, which is used to propagate a helpful semantic feature map and filter redundant informative features irrelevant to instance objects. Finally, we construct branches for feature enhancement to strengthen detailed information to boost the entire feature hierarchy of the network. The experimental results on the Cityscapes dataset manifest that the proposed module outperforms other excellent methods under different evaluation metrics and effectively upgrades the performance of the instance segmentation method.

Estimation in vivo of the chemical composition of the bodies of mature cows

1984 ◽  
Vol 38 (1) ◽  
pp. 33-44 ◽  
Author(s):  
I. A. Wright ◽  
A. J. F. Russel
Keyword(s):  
Body Composition ◽  
Body Fat ◽  
Body Condition ◽  
Body Condition Score ◽  
Deuterium Oxide ◽  
Subcutaneous Fat ◽  
Live Weight ◽  
Eye Muscle ◽  
Ultrasonic Measurements ◽  
Condition Score

ABSTRACTA number of possible indices of body composition (live weight, skeletal size, total body water as estimated by deuterium oxide dilution, blood and red cell volumes as estimated by Evans Blue dilution, ultrasonic measurements of subcutaneous fat depth and eye-muscle area, and body condition scoring) was examined using 73 non-pregnant, non-lactating, mature cows of Hereford × Friesian, Blue-Grey, British Friesian, Galloway and Luing genotypes, ranging in body condition score from 0·75 to 4·5. Direct measurements of body composition in terms of water, fat, protein and ash were made following slaughter.Live weight, deuterium oxide dilution, ultrasonic measurements of subcutaneous fat depth and eyemuscle area, and body condition score were all considered to be potentially useful predictors of body composition. Combinations of techniques offered better predictions than did any single index. Using a combination of measurements it was possible to predict body fat and protein with a residual s.d. of 13·1 kg and 3·15 kg respectively. Breed differences in the partition of fat among the main adipose tissue depots necessitated the development of specific prediction equations for body fat based on condition score and subcutaneous fat depth for different breeds. Equations remain to be developed for predicting body composition in cows in different physiological states.

The effect of body condition and previous nutrition on the herbage intakes of ewes grazing autumn pastures at two sward heights

1994 ◽  
Vol 58 (2) ◽  
pp. 231-235 ◽  
Author(s):  
A. M. Sibbald ◽  
W. G. Kerr
Keyword(s):  
Body Condition ◽  
Body Condition Score ◽  
Day Treatment ◽  
Live Weight ◽  
Feeding Level ◽  
High Body ◽  
Sward Height ◽  
Herbage Intake ◽  
Condition Score ◽  
Grazing Period

AbstractTo examine the effects of body condition and previous nutrition on the herbage intake of ewes grazing swards of different heights in autumn, 96 Scottish Blackface X Border Leicester ewes with a wide range of body condition (score 1·75 to 3·50), were initially housed and given 50 g dry matter (DM) per kg metabolic live weight (M)0·75 per day (treatment L) or 95 g DM per kg M0·75 per day (treatment H) of a pelleted dried grass diet (11·6 MJ metabolizable energy per kg DM) for 6 weeks after weaning in July. The H ewes gained more live weight (9·0 v. 2·7 kg) and body condition score (0·39 v. 0·17) than the L ewes. Half the animals from each treatment were then allocated to each of two ryegrass pastures with a sward height of 5 cm (LS) or 10 cm (HS) for a further 6-week grazing period. During the grazing period there was no significant effect of indoor feeding level on herbage intake, but the L ewes gained more live weight (6·4 v. 5·0 kg) than the H ewes. On the HS, compared with the LS sward, mean herbage intakes were higher (70·0 v. 60·5 g DM per kg M0·75) as were gains in live weight and condition score (7·9 v. 3·4 kg; 0·18 v. 0·0). There were no interactions between the effects of sward height and previous feeding level on herbage intake. Ewes in low body condition (< 2·5) at the start of the grazing period ingested the same amount of herbage on both swards (70·3 g DM per kg M0·75) whereas ewes in high body condition (> 2·5) ingested more (67·0 v. 51·6 g DM per kg M0·75) on the HS compared with the LS sward. The responses of ewes in low and high body condition to different sward heights are discussed in relation to appetite drive and aspects of grazing behaviour.

Prediction of energy status in first lactation dairy heifers

1985 ◽  
Vol 41 (2) ◽  
pp. 167-175 ◽  
Author(s):  
M. J. Ducker ◽  
Rosemary A. Haggett ◽  
W. J. Fisher ◽  
S. V. Morant
Keyword(s):  
Energy Balance ◽  
Confidence Interval ◽  
Body Condition ◽  
Weight Change ◽  
Body Condition Score ◽  
Live Weight ◽  
Blood Metabolites ◽  
Energy Status ◽  
Dairy Heifers ◽  
Condition Score

ABSTRACTData from a large controlled experiment to investigate the effect of level of nutrition on reproductive performance were used to assess the value of production and blood measures as indicators of energy status in lactating dairy heifers. Live-weight change showed the strongest and most consistent relationship to mean energy balance (the difference between metabolizable energy intake and that used for milk production and maintenance) (P < 0·01 to P < 0·001). Body-condition score at a particular time was more closely related to mean energy balance in the preceding 4-week period (P < 0·05) than current energy balance. There was also a lag in the relationship between energy balance and live-weight change and mean body-condition score. Ultrasonic back fat measurements were significantly correlated with both loin and tailhead body-condition score (P < 0·001) but were more strongly related to mean energy balance in the preceding period (P < 0·05 to P < 0·001) than the body-condition scores.Blood samples were taken from all heifers 2 weeks before calving and 1, 5, 9, 13 and 18 weeks after calving and were analysed for 13 constituents. Concentrations of blood metabolites did not show consistently strong correlations with mean energy balance. The only blood metabolite to be measurably affected by the nutritional treatments applied in lactation was β-hydroxybutyrate.At best, combinations of production measures and blood metabolites were only able to predict the mean daily energy balance with a 95% confidence interval of ±20 MJ for an individual animal although this confidence interval was reduced to ±3 MJ for 100 animals.

scholarly journals Precision Betacarotene Supplementation Enhanced Ovarian Function and the LH Release Pattern in Yearling Crossbred Anestrous Goats

Animals ◽  
2020 ◽  
Vol 10 (4) ◽  
pp. 659
Author(s):  
Noé M. Lopez-Flores ◽  
César A. Meza-Herrera ◽  
Carlos Perez-Marin ◽  
Dominique Blache ◽  
Gerardo Arellano-Rodríguez ◽  
...  
Keyword(s):  
Ovarian Function ◽  
Body Condition Score ◽  
Area Under The Curve ◽  
Live Weight ◽  
Beta Carotene ◽  
Ovarian Activity ◽  
Release Pattern ◽  
Time Treatment ◽  
Condition Score ◽  
Lh Release

The possible out-of-season effect of beta-carotene supplementation on ovulation rate (OR), antral follicles (AFN), and total ovarian activity (TOA = OR + AFN) as related to the LH release pattern in yearling anestrous goats was evaluated. In late April, Alpine-Saanen-Nubian x Criollo goats (n = 22, 26 N) were randomly allotted to: (1) Beta-carotene (BETA; n = 10, orally supplemented with 50 mg/goat/d; 36.4 ± 1.07 kg live weight (LW), 3.5 ± 0.20 units, body condition score (BCS) or (2) Non-supplemented (CONT; n = 12, 35.2 ± 1.07 kg LW, 3.4 ± 0.2 units BCS). Upon estrus synchronization, an intensive blood sampling (6 h × 15 min) was accomplished in May for LH quantifications; response variables included (pulsatility-PULSE, time to first pulse-TTFP, amplitude-AMPL, nadir-NAD and area under the curve-AUC). Thereafter, an ultrasonography scanning was completed to assess OR and AFN. The Munro algorithm was used to quantify LH pulsatility; if significant effects of time, treatment or interaction were identified, data were compared across time. Neither LW nor BCS (p > 0.05) or even the LH (p > 0.05); PULSE (4.1 ± 0.9 pulses/6 h), NAD (0.47 ± 0.13 ng) and AUC (51.7 ± 18.6 units) differed between treatments. Nonetheless, OR (1.57 vs. 0.87 ± 0.18 units) and TOA (3.44 vs. 1.87 ± 0.45 units) escorted by a reduced TTFP (33 vs. 126 ± 31.9 min) and an increased AMPL (0.55 vs. 0.24 ± 0.9 ng), favored to the BETA supplemented group (p < 0.05), possibly through a GnRH-LH enhanced pathway and(or) a direct effect at ovarian level. Results are relevant to speed-up the out-of-season reproductive outcomes in goats while may embrace translational applications.

scholarly journals TISNet-Enhanced Fully Convolutional Network with Encoder-Decoder Structure for Tongue Image Segmentation in Traditional Chinese Medicine

2020 ◽  
Vol 2020 ◽  
pp. 1-13
Author(s):  
Xiaodong Huang ◽  
Hui Zhang ◽  
Li Zhuo ◽  
Xiaoguang Li ◽  
Jing Zhang
Keyword(s):  
Image Segmentation ◽  
Receptive Field ◽  
Positional Information ◽  
Dice Similarity Coefficient ◽  
Segmentation Method ◽  
Feature Maps ◽  
Convolutional Network ◽  
Fully Convolutional Network ◽  
Segmentation Methods ◽  
Feature Pyramid

Extracting the tongue body accurately from a digital tongue image is a challenge for automated tongue diagnoses, as the blurred edge of the tongue body, interference of pathological details, and the huge difference in the size and shape of the tongue. In this study, an automated tongue image segmentation method using enhanced fully convolutional network with encoder-decoder structure was presented. In the frame of the proposed network, the deep residual network was adopted as an encoder to obtain dense feature maps, and a Receptive Field Block was assembled behind the encoder. Receptive Field Block can capture adequate global contextual prior because of its structure of the multibranch convolution layers with varying kernels. Moreover, the Feature Pyramid Network was used as a decoder to fuse multiscale feature maps for gathering sufficient positional information to recover the clear contour of the tongue body. The quantitative evaluation of the segmentation results of 300 tongue images from the SIPL-tongue dataset showed that the average Hausdorff Distance, average Symmetric Mean Absolute Surface Distance, average Dice Similarity Coefficient, average precision, average sensitivity, and average specificity were 11.2963, 3.4737, 97.26%, 95.66%, 98.97%, and 98.68%, respectively. The proposed method achieved the best performance compared with the other four deep-learning-based segmentation methods (including SegNet, FCN, PSPNet, and DeepLab v3+). There were also similar results on the HIT-tongue dataset. The experimental results demonstrated that the proposed method can achieve accurate tongue image segmentation and meet the practical requirements of automated tongue diagnoses.

Sign in / Sign up

Export Citation Format

Share Document