Imaging Analytics University Holiday Imaging

Our holiday imaging emails were created to show how our skills could be used to see holiday-related items in a unique and unusual way. They are meant to both entertain and help researchers, product managers, and clinical trial managers think “outside the box” as they consider new ways to analyze and document their work.

 

Christmas 2016Christmas, 2016

This year for Christmas, the ImageIQ Imaging Experts took on the Mannequin Challenge to make a point: 2017 will be a year of rapid change, and that standing still is the least effective course of action. The Mannequin Challenge is a viral Internet video trend where people remain frozen-in-time like mannequins while moving camera films them, usually with the song “Black Beatles” playing in the background. Well, we started out with Black Beatles, but after all, it was the holidays, so we merged into a funky holiday tune. Change is inevitable, and that includes clinical trial imaging and management methodologies. Software-generated clinical trial imaging data can be used to successfully communicate your product’s efficacy story. Starting with fine-tuned imaging protocols, ImageIQ’s expertise lies in helping you tell your story through the right imaging modalities, and with the most objective, quantitative image analysis data possible. All of this, of course, empowered by a truly unique cloud-based imaging-enabled EDC software to make life easier (and more compliant) for you and your clinical trial sites.


Thanksgiving 2016Thanksgiving, 2016

This time of year is also ideal for reflecting on our many blessings, and to say “Thank You” to all our customers for allowing ImageIQ to play a role in supporting their science and research for the betterment of mankind. For this we are truly grateful and humbled. This Thanksgiving, we leveraged imaging and image analysis to automatically detect and measure nitrogen content in corn plants by segmenting various corn leaf characteristics. This methodology allows farmers to know which (and when) crops need more (or less) fertilizer. By optimizing their crop management regimen, farmers can take a “sniper,“ rather than “shotgun,” approach to fertilizer application, and in turn save money and protect the environment by reducing localized chemical run-off.


 

Happy Halloween 2016Halloween, 2016

This year for Halloween, we drew a parallel between spider webs and lung perfusion image analysis. To the naked eye, and even a trained observer, it can be difficult to detect and measure subtle differences within a longitudinal image data set. In the world of clinical trials, the ability to accurately quantify and document those differences – visually and with data – is crucial for FDA approval. See the enhanced data that image analysis can deliver!


 

Fourth of JulyFourth of July, 2016

This year on the 4th of July we focused on safety and showed how a simple sparkler is capable of significant burns.  For our analysis, we used a ham hock as a proxy for human skin to show its burn potential. Top left image is 3D scan of a ham hock without a burn acquired using a 3D surface scanning system attached to an Ipad mini (part of the ImageIQ EDC Derm Platform). The top center image is a 3D scan of the same ham hock with a sparkler burn (red arrow). Bottom images are 3D surface scans acquired using a FARO arm, left image is the native scan, the center image includes the segmented burn volume (green), the right image includes a pseudocolored representation of the burn depth (look-up table scale to the right of the image). The top right graph shows the max burn depth across the length of the burn (blue arrow shows the direction of the analysis).



Memorial Day EsendMemorial Day, 2016

This Memorial Day, the subject of our image analysis honors those who gave up their loved ones for our freedom. We quantified the size and intensity with which our eternal remembrance burns by analyzing this video of the eternal flame. We dissected the video into 400+ frames, converted the original color video to grayscale, and applied a series of spectral and intensity-based filters resulting in a consistent intensity gradient for flame segmentation. Morphology constraint filters were then applied, and segmented flame area was quantified, output and plotted versus time. Mask and area pseudo-colored overlays (right) were co-registered to the raw video (left) to visualize the quantitative data and highlight the dynamic flame boundary (blue) and localized flame intensity (e.g., red corresponding to greater intensity).


St. Patrick's Day 2016April Fools Day, 2016

2016 was our first entry into April Fools day holiday imaging, and being scientists, engineers and others who appreciate our favorite series of movies since childhood, we just had to do it. Take a moment and read about our work in analyzing Midi-chlorian images.


St. Patrick's Day 2016St. Patrick's Day, 2016

They’re Always After Our Lucky Charms.

This year for our St. Patrick’s Day email, the goal was to develop an automated algorithm to analyze the number and distribution of marshmallows and toasted oat pieces within a box Lucky Charms breakfast cereal, and to evaluate the mixed population relative to where a sample was taken from the cereal box. In other words, to maximize marshmallow content in your bowl, should you aim to pour from the top, middle, or bottom of the box?


Valentines 2016Valentines, 2016

This Valentine's Day, we philosophized about where love resides, and used our new ImageQuantify.com platform to have some fun with that question. We set out to quickly and accurately analyze cardiac muscle fiber histological cross sections, extracting the number of fibrils as well as each fibril’s average and individual area, aspect ratio, roundness, perimeter length, and diameter. The quantification of muscle fiber size and shape properties provides researchers valuable insight into many metabolic and cardiovascular disorders. Our unique IQbot filters an image and performs morphological operations to create masks of the image background, connective tissue, and an applied reticulin stain.


Christmas 2015Christmas, 2015

Going green: for the 2014 holiday season, we sought to analyze some common festive greens. We imaged a branch of a Norway spruce and Fraser fir using a micro-CT scanner to identify the number of branches, the average branch thickness, and average branch length. To perform the analysis, we first applied a 3D thinning algorithm to create a skeleton of the CT volume object. We then applied a nodal analysis to quantify the average branch thickness and branch length for both branches. All image analysis routines were full automated and objective.

 


Thanksgiving 2015Thanksgiving, 2015

This Thanksgiving we looked at just one news-worthy aspect of the Thanksgiving tradition of watching American football games: how much difference does a deflated football really make? 3D surface scans of a regulation football were acquired using an infrared laser scanning system (FARO Arm). The football was imaged fully inflated (brown surface in movie) and then deflated to 3 lower psi levels (gray surface is at lowest psi).


Halloween 2015Halloween, 2015

Nothing is spookier than using imaging in your clinical trials, and not being able to generate the best possible data – that’s why our clients appreciate our work. This Halloween, our skull is a surface scan (STL) generated using a Faro laser scanning system. The eyes were generated with a T2-weighted MRI scan, and subsequently co-registered and fused with the skull volume. Boo!


Memorial Day 2015Fouth of July, 2015

The smell of grilling burgers always seems to be a part of any July 4th celebration. As part of this celebration, ImageIQ analyzed the composition of a burger. To perform the analysis, a Micro-CT scan of the burger was acquired. Then, a region-of-interest (ROI) spline was automatically generated around each axial Micro-CT slice in the image volume to segment the burger and bun. Then, different Hounsfield threshold values were applied to the burger volume to identify and segment the fat content (red), unknown high Hounsfield hard nodules (green) and the burger/meat content.

 


Memorial Day 2015Memorial Day, 2015

As part of Memorial Day, we applied our image analysis to an image of a section of Arlington National cemetery counting the number of headstones in the picture’s field-of-view. To automatically identify and count each headstone, the spectral reflection on the headstone was isolated using a series of color channel operations in the RGB (red, green, blue) color space and morphological filters. An overlay image was generated to visually validate the results with a dot placed on every automatically segmented headstone.


St. Patrick's Day 2015St. Patrick's Day, 2015

Erin go Bragh! Getting into the St. Patrick’s Day festivities, we sought to analyze a pot of gold. The analysis started off by acquiring a CT scan of the pot of gold and then applied a number of automated image segmentation and classification methods to enumerate the number and type of coins. On each CT frame, the cross section of each coin was segmented. Then, the coins were classified using a connected components algorithm to convert the 2D coin segments into 3D coins. Last, the 3D volumetric coins were binned and color coded based upon the coin volume.


Christmas 2014Christmas, 2014

Going green: for the 2014 holiday season, we sought to analyze some common festive greens. We imaged a branch of a Norway spruce and Fraser fir using a micro-CT scanner to identify the number of branches, the average branch thickness, and average branch length. To perform the analysis, we first applied a 3D thinning algorithm to create a skeleton of the CT volume object. We then applied a nodal analysis to quantify the average branch thickness and branch length for both branches. All image analysis routines were full automated and objective.


Thanksgiving 2014Thanksgiving, 2014

This Thanksgiving, we really wanted to delve into life in a turkey yard. With our investigative video, we conducted a semi-automated analysis investigating the change in area between a puffed out and non-puffed out turkey. It was performed by fitting a region-of-interest (ROI) around each turkey using local intensity variations to track the contour boundary. The outline of each ROI was superimposed onto the original image and the respective areas were measured. The outputted areas were normalized relative to each other using key landmarks within the image to account for size changes due to depth perception. As scientists, we wanted further data. The turkeys wanted nothing to do with it.


Halloween 2014Halloween, 2014

This Halloween we wanted to show that data that can’t be measured by the human eye can be accurately measured with image analytics. We analyzed a playful pumpkin seed spitting competition, and quantify each contestant’s spitting power, we wrote an algorithm that measured each contestant’s longest seed time of flight, horizontal distance, and total in-flight distance traveled. The algorithm first segmented the seeds on each movie frame and then tracked each pumpkin seed across the entire movie. We validated our results by creating an overlay movie with the segmented pumpkin seeds marked in red and each pumpkin seed’s flight tracked in a different color.


4th of July 2014Fourth of July, 2014

In honor of July 4th, we analyzed a time-lapse movie of a sequin laced American flag as the complete flag came into the field-of-view. For each movie frame, we automatically segmented and quantified the number of red and blue sequins on the flag using the RGB (red, green, blue) color channels in the movie. We outputted the total red and blue sequin counts per frame on the American flag to the integrated overlay graph.


Valentines Day 2014St. Patrick's Day, 2014

We’re always looking for a bit o’ good luck, so this St. Patrick’s Day, we scanned a four-leafed clover on canning on ImageIQ’s MicroCT at 20um resolution. Each leaf was separately segmented and analyzed for surface area and volume. Luck was undetected ;-)

 


Valentines Day 2014Valentines Day, 2014

This Valentine’s Day we focused on the miracle of the human heart by showing a single beating heart cell. The video is an in vitro culture of cardiomyocytes with one cell (in center) contracting. The red overlay shows the displacement area from frame to frame which is also reflected in the graph (displacement area vs time (0.5 sec per frame 50 sec total acquisition).

 


Christmas 2013Christmas, 2013

This year we sponsored a contest to “illuminate” our image analytics skills. At Image IQ this season, we developed an automated image algorithm to analyze the number of distinct red, green, blue, and yellow lights and the perceived overall brightness for each of the decorative colors. To validate our results, we generated an overlay image with colored dots representing each identified light and its perceived color. We also generated a second overlay image that measured the perceived brightness of particular color palettes in the image. This metric was measured as the total area for each identified color as a percentage of the total image area.


Thanksgiving 2013Thanksgiving, 2013

At Thanksgiving 2013, we wanted to communicate our gratitude for our customers for letting us play a small role in bringing their incredible healthcare innovations to those in medical need. We analyzed a time-lapse movie of an autumn tree measuring the change in foliage as we move into the winter months. For each frame in the movie, we automatically segmented the tree's foliage performing a series of color channel operations in the RGB (red, green, blue) color space. Normalizing the foliage to the beginning of the autumn season, we calculated the fraction of remaining foliage. In addition, we calculated the average foliage color for each frame using the hue channel from the HSV (hue, saturation, and value) color space as the tree changed from red to orange to yellow throughout the season.


Halloween 2013Halloween, 2013

Our “No Need to Lose Your Gourd” idea involved a MicroCT scan of a gourd to analyze the gourd volume, pulp volume, and number of seeds. The seeds, gourd skin, and pulp region were segmented on each axial slice of the CT volume using a series of spectral and morphological filters. The number of seeds were calculated by using a connected components algorithm to convert the 2D seed regions-of-interests (ROIs) into 3D seed objects.


Fourth of July 20134th of July Firecrackers, 2013

Getting into the 4th of July holiday festivities, we performed a particle analysis of an SLR video slow motion firecracker explosion. We automatically detected and segmented the particles within each image frame to create an overlay movie with the particles pseudo-colored in green. We quantitatively tracked the mean particle distance from the firecracker showing the radial expansion of particles before falling to the ground. Watch the video!


Memorial Day SendMemorial Day, 2013

We thought the best way to memorialize the incredible sacrifices of our armed forces was to apply the same expertise we leverage every day to analyze biomedical images. Watch the video!

 


Valentines Bumper Sticker Holiday ImagingValentines Day, 2013

Nelson Mandela said that a good heart and a good head are a formidable combination. For Valentines day, we looked at the human heart in a way that showed all its splendor. The love inside it, even for our talented scientists, cannot be imaged. Yet.

 


Holiday LightsChristmas, 2012

As we drive through our neighborhood every holiday season, we see a large variety of light displays: some with lawn ornaments, others with blinding high wattage electric bills, and others with carefully segregated color schemes. However, regardless of the display, we all have an opinion with specific displays we purposely drive by each season and others we create an artificial detour to avoid. At Image IQ this season, we developed an completely automated image analysis algorithm to help predict what light displays people enjoy the most. For each analyzed display, we automatically identified the number of distinct red, green, blue, and yellow lights and the perceived overall brightness for each of these colors. To validate our results, we generated an overlay image with colored dots representing each identified light and its perceived color. We also generated a second overlay image that measured the perceived brightness of particular color palettes in the image. This metric was measured as the total area for each identified color as a percentage of the total image area.


Thanksgiving imaging esendThanksgiving, 2012

Does price make a difference? Maybe only in your wallet. We took two 12 ½ pound turkeys to compare the value of a $16.25 price differential. The two fresh turkeys were scanned on a CT scanner with the protocol optimized to equally distinguish soft tissue and bone. The volumes were rendered and key anatomical regions-of-interest (ROI) were automatically segmented. The volume of each ROI was calculated and tabulated. Click to see the result!


Halloween imaging esendHalloween, 2012

We invited our friends to go “Pumpkin Picking with ImageIQ” — the photo tells a story of a child’s dilemma of picking out the best pumpkin from a large group. We conducted a semi-automated analysis of the pumpkin size by fitting an elliptical region-of-interest around each pumpkin. The outline of each pumpkin’s “best-fit” ellipse was superimposed onto the original image and its respective output metrics including the pumpkin’s area, aspect ratio, and max diameter were generated. Further statistical analysis could be performed on the dataset including a binned histogram of the pumpkin’s size distribution.


Fourth of July, 2012

Our “Celebrate the Little Things in Life” sparkler analysis showed our video capabilities, with a Digital SLR at 60 frames per second video. Analysis tracked the origin and path of each individual spark at it emerged from the core. A second step analysis then pseudo-colored those sparks based on the time that they appeared. The rate of spark production was calculated and plotted as a function of time.


Image IQ Holiday Imaging Memorial Day, 2012

With a more serious tone, we honored our military veterans. A military flag was microscopically imaged and analyzed. The threads that comprise the flag were separated into horizontal and vertical components and the thread count in each direction was determined. Threads were applied with false color to highlight the different directions and ability of the analysis software to separate the threads.


Image IQ Holiday Imaging St. Patrick's Day, 2012

“The Blarney Stone and the FDA” theme was creatively done with a MicroCT of porous stone. Analysis included the identification of the number of holes found within the interior of the stone that were filled with low density material, most likely air. Dense material volume was also calculated and compared to the entire volume of the sample.


Image IQ Holiday Imaging Valentines Day, 2012

Our “Forrest Gump Was Wrong” email was created with a MicroCT of individual pieces of chocolate, with segmentation analysis performed to determine composition.


Image IQ Holiday Imaging New Years Day, 2012

Our “Have a Fruitful New Year” theme was created via a MicroCT of a fruitcake followed by a segmentation analysis to separate the different types of materials and volumetric analysis of each type of material. The pseudo coloring was interested to help aid in the visualization.