On-line learning of mutually orthogonal subspaces for face recognition by image sets.

We address the problem of face recognition by matching image sets. Each set of face images is represented by a subspace (or linear manifold) and recognition is carried out by subspace-to-subspace matching. In this paper, 1) a new discriminative method that maximises orthogonality between subspaces is proposed. The method improves the discrimination power of the subspace angle based face recognition method by maximizing the angles between different classes. 2) We propose a method for on-line updating the discriminative subspaces as a mechanism for continuously improving recognition accuracy. 3) A further enhancement called locally orthogonal subspace method is presented to maximise the orthogonality between competing classes. Experiments using 700 face image sets have shown that the proposed method outperforms relevant prior art and effectively boosts its accuracy by online learning. It is shown that the method for online learning delivers the same solution as the batch computation at far lower computational cost and the locally orthogonal method exhibits improved accuracy. We also demonstrate the merit of the proposed face recognition method on portal scenarios of multiple biometric grand challenge.

Suitability of rich-fish culture under mono and polyculture systems in the boro rice ecosystem

An experiment was conducted in farmers’ fields under Paikgacha thana, Khulna to study the suitability of integrated rice-cum-fish culture. Three treatments namely T1 (Puntius gonionotus), T2 (Puntius gonionotus and Cyprinus carpio) and T3 ( Cyprinus carpio) were included for the study each having three replicates. The fish were stocked at a density of 3750/ha in all the rice plots. The physicochemical parameters of water viz., water depth, temperature, dissolved oxygen, pH, salinity, nitrate and phosphate etc. recorded during the study period were found within optimum range. Of the two cultured species C. carpio attained the highest average individual weight (160g) and survival (81.06%). With respect to biomass and income, highest average fish production and net profit per hectare (306.74kg and Tk. 8177.91) were obtained in T2 and the lowest (184.17kg and Tk. 2049.41) obtained in T1 and a significant variation (p<0.05) in fish production was observed among the treatments while for rice production, it was insignificant. The cost benefit ratio of fish production found were 1:1.29, 1:2.14 and 1:1.90 for T1, T2 and T3, respectively.

Production of Thai sharpunti (Puntius gonionotus Bleeker) in polyculture with carps using low-cost feed

The effectiveness of duckweed and rice bran as a low cost supplementary feed was compared through a six months production trial of rajpunti (Barbodes gonionotus) with carps. Six earthen ponds of 360 m² each were used for the trial. Three ponds received duckweed, while the other three received rice bran as supplementary feed. Fish biomass after six months of rearing increased to an average of 2,056 kg/ha in ponds which received duckweed and 2,056 kg/ha in rice bran treated ponds. The net profit with duckweed and rice bran worked out to Tk. 69.752 and Tk. 73.480 kg/ha, respectively. This study revealed that duckweed is a low cost supplementary feed, particularly for farmers with limited income.

Observation on the production performance of Penaeus monodon with Liza parsia under different cropping system

Study was conducted in six ponds each with an area of 0.1 ha in the pond complex of Brackishwater Station, Paikgacha, Khulna from February to October '96, to find out the variation of production rate in two culture system viz., single and double crop of P. monodon with L. parsia. In treatment T1 wild fry of P. monodon (0.006g) and L. parsia (0.20g) collected from nearby river were stocked at a rate of 40,000 and 10,000/ha, respectively, for a culture period of 120 days. In treatment T2, the rate was 20,000/ha for bagda fry in 1st and 2nd crop each and 10,000 for parsia fry/ha for an extended period of 225 days. The highest survivability and growth of P. monodon and L. parsia were 57.08% (1st crop of T2) and 75.26% (T2), and 27.08g (1st crop of T2) and 47.78g (T2), respectively with a significant variations (P>0.05) with other treatment. The net profit (Tk. 93,134) and cost benefit ratio of 1:1.76 were also found higher in T2.

Effect of high and low cost brood feeds on gonado-somatic index and fecundity of freshwater prawn Macrobrachium rogenbergii de Man

Studies were conducted to observe the effects of different types of feeds on the gonado-somatic index (GSI) and fecundity of freshwater prawn Macrobrachium rogenbergii. Three different treatments (T1 T2 and T3) were designed with three types of feed as follows: (i) Saudi-Bangla Prawn feed 100% - T1 (ii) SaudiBangla prawn feed 50%+ local feed 50%- T2 and (iii) local feed 100%- T3. The results showed that the average value of gonado-somatic index (GSI) was 14.39, 14.35 and 14.36 and the average fecundity of M. rogenbergiiwas 99,741, 98,125 and 97,911 in T1 T2 and T3 respectively. No significant difference (p>00.5) was between gonado-somatic indices (GSI) and fecundities of M. rogenbergii among different feeding trails. The price of Saudi-Bangla prawn feed was very high (Tk. 23/kg) than the local feed (Tk. 14/kg). So, use of local feed was recommended for M. rogenbergii brood rearing.

Growth and yield of GIFT (Oreochromis niloticus) and Thai silver barb (Barbodes gonionotus Bleeker) in rice fields and their effects on the yield of rice

(GIFT) (Oreochromis niloticus) and Silver barb (Barbodes gonionotus) in rice fields and their effects on the yield of rice was carried out in nine experimental rice plots. Three treatments viz., treatment-1 with 0. niloticus (T1), treatment-2 with B. gonionotus (T2) and treatment- 3 was kept as control (T3, without fish) were used in this study. Fertilizers such as, Urea (178 kg ha-1), T.S.P (125 kg ha-1) and M.P. (67 kg ha-1) were applied in each treatment. The fishes were stocked @ 6250 ha·1 and the experiment was continued for a period of 107 days. The values of water quality parameters such as, water temperature, dissolved oxygen, pH and chlorophyll-a were found within suitable level. Between the two species, higher specific growth rate was recorded in 0. niloticus than that of B. gonionotus. But B. gonionotus showed much higher survival (72%) than that of 0. niloticus (35%). Similar to survival, higher production (244 kg ha-1) and income (Tk. 6399 ha-1) were recorded in B. gonionotus than those of 0. niloticus (142.8 kg ha'1 and Tk. 2137 ha-1). Significant differences (p

A comparative study on the effect of commercial fish feeds on the growth of Thai pangas, Pangasius hypophthalmus

A 70 day long experiment was carried out to evaluate three commercial pangas feeds available in Bangladesh viz. Quality Feeds Ltd. (QF), Aftab Bohumukhi Farm Ltd. (ABF) and Saudi-Bangla Fish Feed Ltd. (SBFF) (designated as treatment_T1 T2 and T3) on the growth of Pangas, Pangasius hypophthalmus. Each treatment had two replicates using six experimental ponds of each 80m2 size. The stocking density was 250 fish/80m2. At the beginning, the fish were fed twice daily at 15% of their body weight which was gradually reduced to 10 and 6%, respectively for the rest of the period. The ranges of water quality parameters observed were: temperature 29.0°C- ~5.1 °C, pH 6.85 - 8.52, DO 1.71 - 7.65 mg/1 and transparency or Secchi depth 14.5 - 30.0 em. The mean weight gain of fish was significantly (~<0.05) higher in T3 followed by T2 and T. The specific growth rate (SGR) ranged. between 4.09 and 5.06, feed conversion ratio (FCR) values between 1.54 and 1.61 with treatment T1 showing the lowest FCR. Protein efficiency ratio (PER) values ranged between 2.03 and 2.11. The survival of fish varied between 90.4 and 91.6%. The significantly (P<0.05) highest production of fish (kg/ha/70 days) and profit (Tk/ha/70 days) was observe in T3 (SBFF) followed by T2 (ABF) and T1 (QF), respectively. The result of the study showed that on the basis of nutritive value and growth performance of pangas, feed from Saudi-Bangla Fish Feed Ltd. is the best.

An empirical study on income and efficiency of pond fish and nursery fish production in some selected areas of Pabna, Bangladesh

The study was designed to determine the costs, returns and relative profitability of pond fish and nursery fish production. In order to attain this objective, a total of 70 producers: 35 producing pond fish and 35 producing nursery fish were selected on the basis of purposive random sampling technique from 6 villages under two Upazilas (Sujanagar and Santhia) of Pabna district. It was estimated that per hectare per year gross cost of pond fish production was Tk 65,918 while gross return and net return were Tk 91,707 and Tk 25,789 respectively. Per hectare per year gross cost of nursery fish production was Tk 87,489 while gross return and net return were Tk 1,39,272 and Tk 51,783 respectively. The findings revealed that nursery fish production was more profitable than pond fish production. Cobb-Douglas production function was applied to realize the specific effect of the factors on pond fish and nursery fish production. It was observed that most of the included variables had significant impact on pond fish and nursery fish production. Out of five variables included in the function, all the variables had positive impact on return from pond fish production but stock value of pond, material cost and pond area had positive impact on return from nursery fish production.