% * This code was used in the following articles: % * [1] Learning 3-D Scene Structure from a Single Still Image, % * Ashutosh Saxena, Min Sun, Andrew Y. Ng, % * In ICCV workshop on 3D Representation for Recognition (3dRR-07), 2007. % * (best paper) % * [2] 3-D Reconstruction from Sparse Views using Monocular Vision, % * Ashutosh Saxena, Min Sun, Andrew Y. Ng, % * In ICCV workshop on Virtual Representations and Modeling % * of Large-scale environments (VRML), 2007. % * [3] 3-D Depth Reconstruction from a Single Still Image, % * Ashutosh Saxena, Sung H. Chung, Andrew Y. Ng. % * International Journal of Computer Vision (IJCV), Aug 2007. % * [6] Learning Depth from Single Monocular Images, % * Ashutosh Saxena, Sung H. Chung, Andrew Y. Ng. % * In Neural Information Processing Systems (NIPS) 18, 2005. % * % * These articles are available at: % * http://make3d.stanford.edu/publications % * % * We request that you cite the papers [1], [3] and [6] in any of % * your reports that uses this code. % * Further, if you use the code in image3dstiching/ (multiple image version), % * then please cite [2]. % * % * If you use the code in third_party/, then PLEASE CITE and follow the % * LICENSE OF THE CORRESPONDING THIRD PARTY CODE. % * % * Finally, this code is for non-commercial use only. For further % * information and to obtain a copy of the license, see % * % * http://make3d.stanford.edu/publications/code % * % * Also, the software distributed under the License is distributed on an % * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either % * express or implied. See the License for the specific language governing % * permissions and limitations under the License. % * % */ function OneShot3dEfficient(ImgPath, OutPutFolder,... taskName,...% taskname will append to the imagename and form the outputname ScratchFolder,... % ScratchFolder ParaFolder,... % All Parameter Folder Flag... % All Flags 1) intermediate storage flag ); if ~isdeployed addpath(genpath('../../LearningCode')); addpath(genpath('../../third_party')); addpath(genpath('../../bin/mex')); end % This function is the speed up version of the OneShot3d % Improvement Log: % 1) speedup segment Mex-file % 2) speedup SparseSample3d Mex-file % 3) eliminate reading image and filterbank calculation multiple times % Input: % ImgPath -- the path include the file name of the image % OutPutFolder -- the path of the output folder % ScratchFolder -- intermediante data storage place (used for learning and debug) % Parameter and Data Setting ========================= % startTime = tic; tic fprintf('Starting with new optimization... '); if nargin < 2 disp('Eror: At least need two input argument needed'); return; elseif nargin < 3 taskName = ''; Flag = []; ScratchFolder = ['/afs/cs/group/reconstruction3d/scratch/IMStorage' ]; ParaFolder = '/afs/cs/group/reconstruction3d/scratch/Para/'; elseif nargin < 4 Flag = []; ScratchFolder = ['/afs/cs/group/reconstruction3d/scratch/IMStorage' ]; ParaFolder = '/afs/cs/group/reconstruction3d/scratch/Para/'; elseif nargin < 5 Flag = []; ParaFolder = '/afs/cs/group/reconstruction3d/scratch/Para/'; elseif nargin < 6 Flag = []; end % yalmiptest % parameter setting filename{1} = ImgPath( ( max( strfind( ImgPath, '/'))+1) :end); % Function that setup the Default Default = SetupDefault_New(... [ strrep(filename{1}, '.jpg', '') '_' taskName],... ParaFolder,... OutPutFolder,... ScratchFolder,... Flag); % disp([ num2str( toc(startTime) ) ' seconds.']); disp([ num2str( toc ) ' seconds.']); % Image loading fprintf('Loading the image... '); img = imread(ImgPath); % imgCameraParameters = exifread(ImgPath); % if false %Default.Flag.DisplayFlag && (any( strcmp(fieldnames(imgCameraParameters),'FocalLength') ) || ... % any( strcmp(fieldnames(imgCameraParameters),'FNumber') ) || ... % any( strcmp(fieldnames(imgCameraParameters),'FocalPlaneXResolution') ) || ... % any( strcmp(fieldnames(imgCameraParameters),'FocalPlaneYResolution') ) ) % FocalPlaneResolutionUnit % disp('This image has known f and/or f/sx '); % end % disp([ num2str( toc(startTime) ) ' seconds.']); disp([ num2str( toc ) ' seconds.']); % *************************************************** % Features =========================================== % 1) Basic Superpixel generation and Sup clean fprintf('Creating Superpixels... '); [MedSup, Sup, Default, SupNeighborTable] = gen_Sup_efficient(Default, img); % disp([ num2str( toc(startTime) ) ' seconds.']); disp([ num2str( toc ) ' seconds.']); % 2) Texture Features and inner multiple Sups generation % load /afs/cs/group/reconstruction3d/scratch/Train400/data/MaskGSky.mat; % load /afs/cs/group/reconstruction3d/scratch/Train400/data/LowResImgIndexSuperpixelSep.mat; % load /afs/cs/group/reconstruction3d/scratch/Train400/data/MedSeg/MediResImgIndexSuperpixelSep1.mat % maskg = maskg{1}; % [TextureFeature TextSup]=GenTextureFeature_InnerMulSup(Default, img, Sup{2}, LowResImgIndexSuperpixelSep{1},... % imresize((MediResImgIndexSuperpixelSep),[Default.TrainVerYSize Default.TrainHoriXSize],'nearest'), 1, maskg); % comment compare with old value different only in 1:34 features since % superpixel changes fprintf('Creating Features and multiple segmentations... '); [TextureFeature TextSup]=GenTextureFeature_InnerMulSup(Default, img, Sup{2}, Sup{1},... imresize((MedSup),[Default.TrainVerYSize Default.TrainHoriXSize],'nearest'), 1);%, maskg); % disp([ num2str( toc(startTime) ) ' seconds.']); disp([ num2str( toc ) ' seconds.']); % 3) Superpixel Features generation % [FeatureSupOld, NeighborListOld] = f_sup_old(Default, % LowResImgIndexSuperpixelSep{1}, MediResImgIndexSuperpixelSep); % old % data comparison % new code using prctile replace cause mean diff of 1e-4 fprintf('Calculating superpixel-shape features... '); [FeatureSup] = f_sup_old(Default, Sup{1}, MedSup, SupNeighborTable); % disp([ num2str( toc(startTime) ) ' seconds.']); disp([ num2str( toc ) ' seconds.']); if Default.Flag.IntermediateStorage save([ ScratchFolder '/' strrep( filename{1},'.jpg','') '_IM.mat' ],'FeatureSup','TextureFeature','Sup','TextSup'); end %************************************SID % if Default.Flag.FeatureStorage % name = [strrep(filename{1}, '.jpg', '') '_' taskName]; % save([ScratchFolder '/' name '.mat'],'TextureFeature','FeatureSup'); % end if Default.Flag.FeaturesOnly return; end % at the end with 36130018 bytes in memory % Name Size Bytes Class % DefaTextureFeatureult 1x1 5596 struct array % DepthPara 0x0 0 char array % FeaPara 1x54 108 char array % FeatureSup 13x808 84032 double array % GroundPara 0x0 0 char array % ImgPath 1x78 156 char array % filename 1 66 cell array % MedSup 1200x900 8640000 double array % NeighborList 5134x2 82144 double array % OutPutFolder 0x0 0 char array % SFeaPara 1x54 108 char array % ScratchFlag 1x1 8 double array % ScratchFolder 0x0 0 char array % SkyPara 0x0 0 char array % Sup 1x3 402888 cell array % TextSup 6x2 1611552 cell array % TextureFeature 1x1 13688896 struct array % VarPara 0x0 0 char array % img 2272x1704x3 11614464 uint8 array % taskName 0x0 0 char array % *************************************************** % Inference ========================================== fprintf('Preparation for the Inference... '); % 1) Generate Ground and Sky mask [ maskg, maskSky] = gen_predicted_GS_efficient(Default, TextureFeature.Abs, FeatureSup); % 2) Clean Sup{1} (1st Scale) according to the sky mask [Sup{1}, SupOri, SupNeighborTable]=CleanedSupNew(Default,Sup{1},maskSky, SupNeighborTable); % 3) Generate predicted (depth:1 Variance:2 ) setup as a row verctor [Predicted]=gen_predicted(Default, TextureFeature.Abs, FeatureSup, [1 2]); if Default.Flag.BeforeInferenceStorage save([ ScratchFolder '/' strrep( filename{1},'.jpg','') '_BInf.mat' ], 'Sup', 'SupOri', 'MedSup', 'Predicted', 'maskg', 'maskSky'); end if Default.Flag.NonInference return; end % disp([ num2str( toc(startTime) ) ' seconds.']); disp([ num2str( toc ) ' seconds.']); fprintf('Starting Inference... '); % 4) Plane Parameter MRF RunCompleteMRF_efficient( Default, img, Predicted, MedSup, Sup, SupOri, TextSup, SupNeighborTable, ... reshape( FeatureSup( TextureFeature.Abs(:,1)), Default.VertYNuDepth, []), ... maskSky, maskg); % disp(['Finished Inference at: ' num2str( toc(startTime) ) ' seconds.']); disp(['Finished Inference at: ' num2str( toc ) ' seconds.']); % 5) output data fprintf('Writing superpixels and image... '); save([ Default.OutPutFolder Default.filename{1} '.mat'],'MedSup'); % 6) Image copy to OutPutFolder %system(['cp ' ImgPath ' ' OutPutFolder Default.filename{1} '.jpg']); copyfile(ImgPath, [OutPutFolder Default.filename{1} '.jpg'],'f'); % disp([ num2str( toc(startTime) ) ' seconds.']); % disp(['Done. Total time taken = ' num2str( toc(startTime) ) ' seconds.'] ); disp([ 'Done. Total time taken = ' num2str( toc ) ' seconds.']); % *************************************************** return;