ImportError: libjasper.so.1: cannot open shared object file: No such file or directory

usage: getcoco17.py [-h] [--data_root_dir DATA_ROOT_DIR]
                    [--category CATEGORY [CATEGORY ...]]
                    [--annotation_file ANNOTATION_FILE]
                    [--resolution RESOLUTION] [--id ID [ID ...]]
                    [--number NUMBER]

Download COCO 2017 image Data

optional arguments:
  -h, --help            show this help message and exit
  --data_root_dir DATA_ROOT_DIR
                        The root directory for your data.
  --category CATEGORY [CATEGORY ...]
                        MS COCO object categories list (--category person dog
                        cat). default value is person
  --annotation_file ANNOTATION_FILE
                        The annotation json file directory of MS COCO object
                        categories list. file name should be
                        instances_val2017.json
  --resolution RESOLUTION
                        The resolution of images you want to transfer. It will
                        be a square image.Default is 0. resolution = 0 will
                        keep original image resolution
  --id ID [ID ...]      The id of images you want to download from MS COCO
                        dataset.Number of images is equal to the number of
                        ids. Masking will base on category.
  --number NUMBER       The total number of images you want to download.

usage: main.py [-h] --data_root_dir DATA_ROOT_DIR
               [--weights_path WEIGHTS_PATH] [--split_num SPLIT_NUM]
               [--net {segcapsr3,segcapsr1,capsbasic,unet,tiramisu}] [--train]
               [--test] [--manip] [--shuffle_data {0,1}] [--aug_data {0,1}]
               [--loss {bce,w_bce,dice,bce_dice,mar,w_mar}] [--batch_size BATCH_SIZE]
               [--initial_lr INITIAL_LR] [--steps_per_epoch STEPS_PER_EPOCH]
               [--epochs EPOCHS] [--patience PATIENCE] [--recon_wei RECON_WEI]
               [--slices SLICES] [--subsamp SUBSAMP] [--stride STRIDE]
               [--verbose {0,1,2}] [--save_raw {0,1}] [--save_seg {0,1}]
               [--save_prefix SAVE_PREFIX] [--thresh_level THRESH_LEVEL]
               [--compute_dice COMPUTE_DICE]
               [--compute_jaccard COMPUTE_JACCARD]
               [--compute_assd COMPUTE_ASSD] [--which_gpus WHICH_GPUS]
               [--gpus GPUS] [--dataset {luna16,mscoco17}]
               [--num_class NUM_CLASS] [--Kfold KFOLD] [--retrain {0,1}]
               [--loglevel LOGLEVEL]

Train on Medical Data or MS COCO dataset

optional arguments:
  -h, --help            show this help message and exit
  --data_root_dir DATA_ROOT_DIR
                        The root directory for your data.
  --weights_path WEIGHTS_PATH
                        /path/to/trained_model.hdf5 from root. Set to "" for
                        none.
  --split_num SPLIT_NUM
                        Which training split to train/test on.
  --net {segcapsr3,segcapsr1,capsbasic,unet,tiramisu}
                        Choose your network.
  --train               Add this flag to enable training.
  --test                Add this flag to enable testing.
  --manip               Add this flag to enable manipulation.
  --shuffle_data {0,1}  Whether or not to shuffle the training data (both per
                        epoch and in slice order.
  --aug_data {0,1}      Whether or not to use data augmentation during
                        training.
  --loss {bce,w_bce,dice,bce_dice,mar,w_mar}
                        Which loss to use. "bce" and "w_bce": unweighted and
                        weighted binary cross entropy, "dice": soft dice
                        coefficient, "bce_dice": binary cross entropy + soft dice coefficient, "mar" and "w_mar": unweighted and
                        weighted margin loss.
  --batch_size BATCH_SIZE
                        Batch size for training/testing.
  --initial_lr INITIAL_LR
                        Initial learning rate for Adam.
  --steps_per_epoch STEPS_PER_EPOCH
                        Number of iterations in an epoch.
  --epochs EPOCHS       Number of epochs for training.
  --patience PATIENCE   Number of patience indicates the criteria of early
                        stop training.If score of metrics do not improve
                        during the patience of epochs, the training will be
                        stopped.
  --recon_wei RECON_WEI
                        If using capsnet: The coefficient (weighting) for the
                        loss of decoder
  --slices SLICES       Number of slices to include for training/testing.
  --subsamp SUBSAMP     Number of slices to skip when forming 3D samples for
                        training. Enter -1 for random subsampling up to 5% of
                        total slices.
  --stride STRIDE       Number of slices to move when generating the next
                        sample.
  --verbose {0,1,2}     Set the verbose value for training. 0: Silent, 1: per
                        iteration, 2: per epoch.
  --save_raw {0,1}      Enter 0 to not save, 1 to save.
  --save_seg {0,1}      Enter 0 to not save, 1 to save.
  --save_prefix SAVE_PREFIX
                        Prefix to append to saved CSV.
  --thresh_level THRESH_LEVEL
                        Enter 0.0 for masking refine by Otsu algorithm. Or set
                        a value for thresholding level of masking. Value
                        should between 0 and 1.
  --compute_dice COMPUTE_DICE
                        0 or 1
  --compute_jaccard COMPUTE_JACCARD
                        0 or 1
  --compute_assd COMPUTE_ASSD
                        0 or 1
  --which_gpus WHICH_GPUS
                        Enter "-2" for CPU only, "-1" for all GPUs available,
                        or a comma separated list of GPU id numbers ex:
                        "0,1,4".
  --gpus GPUS           Number of GPUs you have available for training. If
                        entering specific GPU ids under the --which_gpus arg
                        or if using CPU, then this number will be inferred,
                        else this argument must be included.
  --dataset {luna16,mscoco17}
                        Enter "mscoco17" for COCO dataset, "luna16" for CT
                        images
  --num_class NUM_CLASS
                        Number of classes to segment. Default is 2. If number
                        of classes > 2, the loss function will be softmax
                        entropy and only apply on SegCapsR3** Current version
                        only support binary classification tasks.
  --Kfold KFOLD         Define K value for K-fold cross validate default K =
                        4, K = 1 for over-fitting test
  --retrain {0,1}       Retrain your model based on existing weights. default
                        0 = train your model from scratch, 1 = retrain
                        existing model. The weights file location of the model
                        has to be provided by --weights_path parameter
  --loglevel LOGLEVEL   loglevel 3 = debug, 2 = info, 1 = warning, 4 = error,
                        > 4 =critical

usage: gen_mask.py [-h] [--net {segcapsr3,segcapsr1,capsbasic,unet,tiramisu}]
                   --weights_path WEIGHTS_PATH [--num_class NUM_CLASS]
                   [--which_gpus WHICH_GPUS] [--gpus GPUS]

Mask image by segmentation algorithm

optional arguments:
  -h, --help            show this help message and exit
  --net {segcapsr3,segcapsr1,capsbasic,unet,tiramisu}
                        Choose your network.
  --weights_path WEIGHTS_PATH
                        /path/to/trained_model.hdf5 from root. Set to "" for
                        none.
  --num_class NUM_CLASS
                        Number of classes to segment. Default is 2. If number
                        of classes > 2, the loss function will be softmax
                        entropy and only apply on SegCapsR3** Current version
                        only support binary classification tasks.
  --which_gpus WHICH_GPUS
                        Enter "-2" for CPU only, "-1" for all GPUs available,
                        or a comma separated list of GPU id numbers ex:
                        "0,1,4".
  --gpus GPUS           Number of GPUs you have available for training. If
                        entering specific GPU ids under the --which_gpus arg
                        or if using CPU, then this number will be inferred,
                        else this argument must be included.

----SegCaps-master  (Project folder)
    |
    \-cococrawler (Crawler program folder)
    |   \-annotations (Folder of Microsoft COCO annotation files)
    \-data  (The root folder of program output)
    |   \-imgs (Folder of training and testing images)
    |   \-masks (Folder of training and testing masking data)
    |   \-np_files (Folder to store processed image and mask files in numpy form.)
    |   \-split_lists (Folder for training and testing image splits list)
    |   \-logs (Training logs)
    |   \-plots (Trend diagram for Training period. Only generate after the training completed )
    |   \-figs (Conver image to numpy format, part of images stored for checking)
    |   \-saved_models (All model weights will be stored under this folder)
    |   \-results (Test result images will be stored in this folder)
    |     |
    |     \segcapr3\split_0\final_output
    |                      \raw_output
    |
    \-models (Reference model files: Unet and DenseNet)
    |
    \-segcapsnet (main modules for Capsule nets and SegCaps)
    |
    \-utils (image loader, loss functions, metrics, image augmentation, and thread safe models)
    |
    \-notebook (Some experiment notebooks for reference)
    |
    \-raspberrypi (Raspberry Pi software installation scripts) 
    |
    \-imgs (image file for this readme)

1-1. Porting program from python 2.7 to 3.6 (Jun 11)

1-2. Execute manipulation function. (Jun 11)

1-2. Execute test function on one image without pre-trained weight(Jun 11)

1-3. Execute train function on 3 images. (Jun 12)

1-4. Execute test function on trained model (Jun 12)

1-5. Display original image and result mask image. (Jun 12)

1-6. Identify input image mask format. (Jun 14)

2-1. Identify COCO stuff 2017 as target dataset. (Jun 15)

2-2. Download annotation files for COCO 2017. (Jun 15)

3-1. Generate single class mask on COCO masked image data.(Jun 13)

3-2. Convert the image mask format to background=0, objects=1. (Jun 18)

3-3. Convert the color image to gray scale image (Jun 18)

3-3. Feed the color images with single classes mask to model for training. (Jun 21)

4-1. Modify code to support experiments.(Jun 25)

  4-1-1. Models persistent by version with configuration and dataset. (Jun 26)

  4-1-2. Notebook folder build up to store experiment results.

4-2. Test pipeline (Jun 27)

@article{lalonde2018capsules,
  title={Capsules for Object Segmentation},
  author={LaLonde, Rodney and Bagci, Ulas},
  journal={arXiv preprint arXiv:1804.04241},
  year={2018}
}