Use of Digital Images for Evaluation of Factors Responsible for Visual Preference of Apples by Consumers

This research used digital images to explore some of the factors responsible for consumer preference of visual characteristics of apples (Malus ×domestica Borkh.). The images systematically varied in color and shape (Expt. A: 9 images) and type, shape, and background color (Expt. B: 10 images), while keeping apple size constant. Visual assessments of the apple images were collected from 144 consumers (Expt. A) and 165 consumers (Expt. B) in British Columbia (BC), Nova Scotia (NS), and New Zealand (NZ) using balanced incomplete block designs. Canadian consumers (BC and NS) preferred red apples over green or yellow. NZ consumers liked equally red and green apples, and preferred both to yellow apples. At all locations, consumers in Expt. A significantly preferred round and conical shaped apples to oblong apples. When the combined effects of type, shape, and background color were evaluated, NZ consumers rated the striped, round apples the highest, and least preferred both round and oblong, blush-type apples with yellow backgrounds. NS consumers tended to prefer blush apples regardless of type and background color, and BC consumers were more accepting of a range of apple types, shapes, and background colors. Apple breeding is a challenging task because of the need to screen a large number of seedlings combined with the difficulty of “short-listing” the most relevant attributes for the consumer and their relative importance. This challenge is made even more difficult because apple characteristics do not vary systematically. For example, variables such as size, shape, and color all vary concomitantly from cultivar to cultivar and cannot be easily studied in isolation from one another. With digital imagery, photographic images can be electronically modified to systematically study visual traits in apples. The objective of this work was to examine some of the factors responsible for consumer preferences of visual characteristics of apples using digitally modified images. This research was implemented in a series of two experiments which evaluated apple color and shape (Expt. A), and apple type, shape, and backgroundcolor (Expt. B). Two apple images, A8 and A9 (Table 1) were included in Expt. B as B1 and B2, respectively, so that the mean preference scores for the two experiments could be compared. Materials and Methods Apple image creation. Two apples from each of the following cultivars (‘Delicious’, ‘Gala’, and ‘McIntosh’) were photographed using a 35-mm camera containing Kodak professional slide film. One apple was oriented on its side to obtain a ‘calyx’ view, while the other apple was left standing to obtain a lateral/top view. The film was developed and scanned using Kodak Professional RFS 3570 Film Scanner (Eastman Kodak Co., Rochester N.Y.) with an image size of 3072 × 2048 pixels. Images were then edited using Adobe Photoshop version 5.0 (Adobe Systems, San Jose Calif.) on a Power Macintosh 7300/200 computer. Photographic images were modified to create conical, round and oblong shaped apple images; shapes which were consistent with ‘Gala’, ‘McIntosh’, and ‘Delicious’ apples, respectively. The size of all images was standardized to a diameter of 7.5 cm. This size has been shown to be “ideal” by BC consumers (Hampson & Quamme, 2000) and preferred by German, Dutch, and British consumers (Goedegebure et al., 1990). For each shape, three color images (red, yellow, and green) were digitally created. These images served as templates for the creation of images for striped and blushed apples. Photographic images of ‘Gala’ (stripe) and ‘McIntosh’ (blush) were transformed to the selected shapes (as described above) using MORPH software version 2.5 (Gryphon Software Corp. 1992–94, San Diego, Calif.). Care was taken to match the orientation of the apples, so that the ‘stripes’ or ‘blush’ on the original images were aligned with the contours of the digital images. This ensured that the ‘shadows’ and ‘highlights’ were in the appropriate locations and the resultant transformations were as realistic as possible. However, for the stripe and blush apples, it was only practical to create Consumer preference for new apple cultivars is extremely complex and not well understood. In general, consumer preference is dependent on familiarity, socioeconomic status, age, gender, culture, and social norms (Lyman, 1989; Andani and MacFie, 2000). For fruit and vegetables, appearance plays a major role in consumer assessment of quality (von Alvensleben & Meier, 1990) and purchase intent (Crassweller and Hollender, 1989). Regional apple preferences within Canada have been demonstrated (Cliff et al., 1999) to differ with British Columbian (BC) consumers preferring the appearance of ‘Gala’ and ‘Creston’ apples and Nova Scotian (NS) consumers preferring the appearance of ‘Fiesta’ and ‘Cortland’ apples. In the international marketplace, ‘Granny Smith’ and ‘Newtown’ have been shown to be most preferred in the United States, but not in Asia (Stebbins, 1994). Dalliant-Spinnler et al. (1996) used internal and external preference mapping to relate consumer preference of apple varieties with a profile from a trained sensory panel, thereby identifying the most relevant sensory attributes. Two external attributes (white specks, looks hard) were found to be significantly correlated with consumer preference for southern hemisphere apples. Hampson and Quamme (2000) also studied consumer preference of visual attributes in apples. They found that panellists liked the most common shapes, but that lenticels and russet became unacceptable when they exceeded 1.0 mm or 55% of the fruit diameter, respectively. Received for publication 11 Apr. 2001. Accepted for publication 2 Oct. 2001. Contribution No. 2118. We would like to acknowledge the assistance of Michael Weis, John Hall, Marj King, and Kimberly Dever of the Pacific Agri-Food Research Centre (Summerland, B.C.), Ken McRae, Vivian Agar, Christine Burbidge-Boyd, and Connie Foote of the Atlantic Food and Horticulture Research Centre (Kentville, N.S.) and Dogan Yuksel of the Food Innovation Center (Portland, Ore.). To whom reprint requests shuld be addressed. E-mail address: [email protected] Work conducted at Hort Research Auckland, New